Note: Descriptions are shown in the official language in which they were submitted.
CA 02519617 2007-04-05
A MULTIPLEXER FOR ETHERNET NETWORKS
This application is a division of Canadian Application Serial No.
2,458,694, filed February 25, 2004. The claims of the present application are
directed to an Ethernet multiplexer. However, for the purpose of understanding
the invention, including all objects and features which are inextricably bound-
up
in one and the same inventive concept, the teachings of those features claimed
in the parent Canadian Application Serial No. 2,458,694 are retained herein.
Accordingly, the retention of any such objects or features which may be
more particularly related to the parent application or a separate divisional
thereof should not be regarded as rendering the teachings and claiming
ambiguous or inconsistent with the subject matter defined in the claims of the
divisional application presented herein when seeking to interpret the scope
thereof and the basis in this disclosure for the claims recited herein.
FIELD OF THE INVENTION
The present invention relates to an Ethernet multiplexer for relaying
information transmitted/received between Ethernet networks for
accommodating client lines from client terminals and a plurality of types of
transmission networks for connecting the Ethernet networks to each other has
a GbE_MAC processing circuit, a TDM_MUX/DEMUX circuit, an FE_MAC
processing circuit, and a GBP (Generic Blocking Procedure disclosed in United
States Patent Application Publication No. 2004/0136401 published on July 15,
2004) processing circuit. The Ethernet multiplexer serves as an edge node of
each of the Ethernet networks. The Ethernet multiplexer can also have a
terminator circuit, a TDM_MUX/DEMUX circuit, a MAC processing circuit, and a
GBP delay processing circuit.
1
CA 02519617 2005-09-14
BACKGROUND OF THE INVENTION
The proliferation and advancement of the Internet, intranets, and
portable telephones cause an annual increase in the amount of speech and
data related traffic which flows through networks. In such an environment,
enterprises and service providers are imminently driven to build a wide area
Ethernet network at low cost which can support the ever increasing amount of
traffic.
The wide area Ethernet network is built, for example, by use of an
existing SONET/SDH network as a relay section to interconnect LANs, each
implemented by an Ethernet network in such a manner that they can
communicate with one another. For transmission in such a configuration,
higher-level protocol data transferred on a LAN is encapsulated at an ingress
node, which is the entrance of the wide area Ethernet network, in accordance
with a protocol employed in a relay section, and is decapsulated at an egress
node, which is the exit of the wide area Ethernet network. Techniques for
such transmission in the wide area Ethernet network have already been
brought into practical use, examples of which may be PPP over SONET (IETF
RFC2615 standard), GFP (ITU-T G. 7041 standard), and the like. For example,
when a SONET/SDH network is utilized as a relay section, encapsulated data is
stored in a payload of a frame defined in the SONET/SDH network, and
transferred by cross-connecting communication channels with a band which is
previously set.
In the SONET/SDH network, the frame used therein comprises two
types of overhead fields: SOH (Section Over Head) and POH (Path Over Head)
(see, for example, ITU-T Recommendation G. 707 (October 2000)). SOH is
provided for managing a section which is defined as a transmission system
portion of a transmission medium, while POH is provided for managing a path
2
CA 02519617 2005-09-14
network layer which is independent of the transmission medium. In this way,
the section and path are organized in a layered structure, and a plurality of
paths are multiplexed in a payload, such that a transmission network can be
organized in a layered structure comprised of a transmission medium network
layer and a path network layer. This results in the ability to manage the
designing, maintenance and operation of the network in a layered structure,
thereby providing advanced network services. For example, since each relay
node on a transmission path separately monitors SOH fault information and
POH fault information, it is possible to readily find whether a communication
fault, if any, is occurring in a section between the relay nodes or only on a
particular path, and to identify the location at which the fault is occurring.
However, since the SONET/SDH network has a large number of items to
be monitored using SOH and POH, the maintenance and operation costs tend
to increase. To limit the maintenance and operation costs, less expensive
Ethernet networks represented by Fast Ethemet (hereinafter abbreviated as
"FE") and Gigabit Ethernet (hereinafter abbreviated as "GbE") have been
increasingly employed in wide area Ethernet networks for use as relay sections
thereof. Alternatively, another wide area Ethemet network may include relay
sections which may be implemented by a combination of an existing
SONET/SDH network and an Ethernet network such as a GbE network, such
that the SONET/SDH network is utilized in a certain section, while the GbE
network is utilized in another section.
In the wide area Ethemet network, there are many clients who request
guaranteed transmission bandwidth and communication quality. To meet such
a request, a protection function is needed for switching to a spare line if a
line
fault occurs. Also, aiarm information indicative of the occurrence of a fault
must
be notified to clients placed in communication without fail. Further, since a
3
CA 02519617 2005-09-14
large capacity of data is flowing through lines, a need exists for redundant
transmission paths in relay sections and a reduction in a switching time upon
occurrence of a fault.
In the aforementioned GFP, a client management frame shown in Fig. 1
is defined for transferring an alarm for notifying the occurrence of a fault
such
as a line fault, a failed device, and the like. Fig. 1 is a schematic diagram
showing a format for the client management frame defined in the GFP.
As shown in Fig. 1, in the GFP, a communication partner can be notified
of a fault occurring on a client line (for example, on an Ethernet line) using
a
UPI (User Payload Identifier) field in the client management frame. The UPI
value indicates loss of signal (LOS) when it takes "00000001 " and indicates
link-down on a client line when it takes "10000000. " In a normal state, where
no fault is present, the UPI value is set to a value other than the foregoing
"00000001 " and "1000000. "
The client management frame is transferred to a communication partner
at predetermined intervals while loss of signal or link-down is being
detected.
In different systems, the client management frame may be transferred to the
communication partner as required, or transferred at predetermined intervals
in
a normal state where no fault is present.
A core header field shown in Fig. 1 includes such information as the
source address, destination address, and priority, and the PTI (Payload Type
Identifier) field shows how the frame is used. In Fig. 1, the PTI value is set
to
"100", which indicates that the frame is used as a client management frame.
The PFI (Payload FCS Indicator) field shows whether or not FCS (Frame Check
Sequence) is executed. The FCS is used for detecting transmission errors of
payloads. An EXI (Extension Header Identifier) field stores an identifier of
an
extension header when it is used. In Fig. 1, the PFI value is set to "0" which
4
CA 02519617 2005-09-14
indicates that the FCS is not executed, and the EXI value is set to "0000"
which
indicates that no extension header is used.
Next, a conventional alarm transfer method will be described with
reference to a wide area Ethernet network, as an example, which utilizes a
SONET/SDH network as illustrated in Fig. 2. Fig. 2 is a block diagram
illustrating an exemplary configuration of a conventional wide area Ethernet
network.
The wide area Ethernet network illustrated in Fig. 2 comprises a plurality
of Ethernet networks (two in Fig. 2) which are connected through a plurality
of
relay devices (two in Fig. 2) which make up the SONET/SDH network. The
Ethernet networks are connected to relay devices 203, 204 of the SONET/SDH
network through Ethernet termination devices 202, 205 contained in the
respective Ethernet networks. Each of the Ethernet networks accommodates a
plurality of client terminal devices 201, 206 (one each at the respective ends
in
Fig. 2, hereinafter called the "client terminal"). Ethernet termination
devices
202, 205 each time-division-multiplex data transmitted from the respective
client terminals to generate SONET/SDH frames which are transmitted to relay
devices 203, 204. Upon receipt of a SONET/SDH frame from relay device 203
or 204, the frame is demultiplexed into data for respective client terminals,
and
demultiplexed data are transmitted to the associated client terminals.
For transmitting data from an arbitrary client terminal to an opposing
client terminal, a direction in which the data is transmitted is defined as a
forward direction, and a direction opposite to the forward direction is
defined as
a backward direction. Also, a client terminal which originates data is said to
be
located upstream, while a client terminal which receives the data is called to
be
located downstream.
5
CA 02519617 2005-09-14
Each of devices which relays a communication between client terminals
is collectively called the "relay node," in particular, a node which receives
a
signal from a client line or client circuit is called an "ingress node," and a
node
which delivers a signal to a client line or client circuit is called an
"egress node.
" In the configuration shown in Fig. 2, assuming that the data source is
client
terminal 201, and the data destination is client terminal 206, Ethernet
termination device 202 is in position of the ingress node, while Ethernet
termination device 205 is in position of the egress node. A data flow from a
client node serving as the data source to a client node serving as the data
destination through respective relay nodes is called an "Ethemet path. " Relay
nodes passed by a certain Ethernet path, and their port numbers have been
previously set in the respective relay nodes. During a normal operation, an
Ethernet path will not change to pass different relay nodes from previously
set
ones.
In the configuration as described above, assuming that a fault occurs, for
example, on the client line between client terminal 201 in position of the
data
source and Ethernet termination device 202 to result in loss of signal or
link-down, Ethernet termination device 202, which has detected the fault,
transmits a client management frame indicative of loss of signal or link-down
to
Ethernet termination device 205 which is connected to client terminal 206 of
the
communication partner that is set in an Ethernet path. The client management
frame indicative of loss of signal or link-down is transmitted at
predetermined
intervals while loss of signal or link-down is being detected.
Upon detection of the client management frame, Ethernet termination
device 205 stops deiivering a signal to associated client terminal 206, and
forcefully sets the line to loss of signal or link-down. This state is
maintained at
6
CA 02519617 2005-09-14
all times while Ethernet termination device 205 is detecting the client
management frame indicative of loss of signal or link-down.
In this way, as a fault occurs on a client line, line fault information is
communicated to an opposing client line, so that the client lines appear to be
directly connected to each other without causing client terminals 201, 206 to
be
aware of the existence of the intervening SONET/SDH network. This function
for allowing the client terminal not to be aware of the existence of an
intervening line or circuit is called "link-pass-through."
For providing redundant relay lines in the SONET/SDH network,
SONET/SDH has provided a mechanism for realizing a switching from a failed
line to a spare line within 50 milliseconds upon occurrence of a fault. For
operating this mechanism, an SOH field has one byte each of K1 byte and K2
byte which are communicated to an opposing communication device that
serves as a communication partner (see, for example, ITU-T Recommendation
G. 841 (Qctober 1998)).
In another example of providing redundancy for Ethernet lines wherein
redundant lines are provided between two communication devices, an active
transmission line is set to link-down in response to detection of a fault in
an
active system to notify an opposing communication device of the fault on the
active transmission path. Upon detection of link-down, the opposing
communication device switches the active system to a redundant system (see,
for example, ITU-T Recommendation G. 707 (October 2000)).
The conventional wide area Ethernet network which applies the
aforementioned GFP has the inability to accomplish link-pass-through for
transferring information on a fault on a relay line to a client terminal of a
communication partner, when an Ethemet path is relayed through a plurality of
types of transmission networks, as is the case with a combination of the GbE
7
CA 02519617 2005-09-14
network and SONET/SDH network. This result occurs because a field for use
in transferring an alarm provided in the client management frame defined by
GFP can merely transfer information on a fault which has occurred in a client
line section.
As an example, when relay devices for building the GbE network are
installed between the Ethernet termination devices and the relay devices of
the
SONET/SDH network illustrated in Fig. 2, and a line fault occurs, for example,
between the upstream Ethernet termination device and a relay device in the
GbE network, this fault information cannot be transferred downstream using the
client management frame because the fault information notifies a fault on a
relay line. If the UPI field in the client management frame were used to
transfer
information on loss of signal or link-down caused by a fault on a relay line
in a
manner similar to a fault on a client line, the downstream Ethernet
termination
device could not distinguish a client line fault from a relay line fault. This
will
result in difficulty in identifying the location of the fault. The relay
device in the
GbE network is similar to the Ethernet termination device in that it time-
division
multiplexes data received from a plurality of Ethernet termination devices for
transmission to the SONET/SDH network, and demultiplexes frames received
from the SONET/SDH network into individual data for transmission to
associated Ethemet termination devices.
To solve the above deficiency, in a wide area Ethernet network which
utilizes the GbE network and SONET/SDH network, inherent alarm transfer
systems defined in the GbE network and SONET/SDH network may be used to
sequentially transfer the alarm information to downstream client lines.
For example, in a wide area Ethernet network comprised only of the
SONET/SDH network illustrated in Fig. 2, as a fault occurs between relay
devices within the SONET/SDH network, the fault information is transferred to
a
8
CA 02519617 2005-09-14
downstream Ethernet termination device through a path AIS (Alarm Indication
Signal) alarm defined by SONET/SDH, and also transferred to an upstream
Ethernet termination device through a path RDI (Remote Defect Indication)
alarm defined by SONET/SDH. Ethemet termination device 205 implements
the link-pass-through by setting an Ethernet line to link-down only for those
client terminals which utilize a path for which the path AIS is detected.
Similarly, Ethernet termination device 202 implements the I in k-pass-th rough
by
setting an Ethernet line to link-down only for those client terminals which
utilize
a path for which the path RDI is detected.
Assuming that the alarm transfer systems inherent to the GbE network
and SONET/SDH network are used for sequentially transferring alarm
information to downstream client lines in a wide area Ethernet network, which
has relay devices for building the GbE network, each installed between each of
the Ethernet termination devices and each of relay devices of the SONET/SDH
network as illustrated in Flg. 2, a fault in the GbE section would cause
forced
disconnection of not only a line which connects a failed Ethernet termination
device to a relay device of the GbE network but also lines, each of which
connects a normal relay device of the GbE network multiplexed with the relay
device involved in the fault to a relay device of the SONET/SDH network. As a
result, a normal Ethemet path from another Ethernet termination device which
passes through the relay device of the GbE network is also forcefully set to
link-down.
Conversely, in a wide area Ethernet network which stores a GFP
capsule in a data field of a MAC frame, such as GbE, for transfer, there is
the
inability to organize a transmission network in a layered structure comprised
of
a transmission medium network layer and a path network layer for designing,
maintaining, and operating the network. This results from no field being
9
CA 02519617 2005-09-14
defined in the MAC frame used in the Ethernet network for distinguishing the
transmission medium network layer from the path network layer, thus failing to
manage the section and path in a layered structure, as is done in the
SONET/SDH network.
The MAC frame is comprised of a preamble field (seven octets long); an
SFD (Start Frame Delimiter) field (one octet long) for frame identification; a
destination MAC address (hereinafter abbreviated as "DA") field (six octets
long); a source MAC address (hereinafter abbreviated as "SA") field (six
octets
long); a LENGTH/TYPE field (two octets long) indicative of the frame length or
type; a data field (46 to 1,500 octets long); and an FCS (Frame Check
Sequence) field (four bytes) for CRC operation. Errors can be detected in
received frames by monitoring the FCS field.
However, the MAC frame lacks a field for distinguishing the section from
the path for management, so that even if an error is detected in the CRC
operation with the FCS field, it is not possible to immediately identify
whether
the error is associated with a fault in a transmission medium network layer or
with a fault in an individual path network layer independent of a transmission
medium. For identification, the detected error must be compared with fault
information generated by the preceding and subsequent relay nodes.
Furthermore, in a wide area Ethernet network which includes relay
sections comprising a combination of an existing SONET/SDH network and an
Ethemet network, such as a GbE network, the SONET/SDH network is utilized
in certain sections and the GbE network is utilized in the remaining sections.
When an Ethernet network switching means, used for providing redundancy for
both the SONET/SDH network and Ethemet network, involves forcefully
bringing an active system, which detects an existing fault on a transmission
path, to link-down by identifying the fault on the transmission path to an
CA 02519617 2005-09-14
opposing communication device for switching the system. A relay node, which
serves as a junction of the SONET/SDH network with the Ethernet network,
must be provided with two different types of switching means, i. e. , a
switching
means which uses the K1 byte and K2 byte of the SONET/SDH network, and a
switching means which uses link-down of the Ethernet network, thus resulting
in
an increase in the circuit scale, mounting area, and switching time.
SUMMARY OF THE INVENTION
It is an object of the present invention to provide an alarm transfer
method for use in a wide area Ethernet network which utilizes a plurality of
transmission networks such as the GbE network, SONET/SDH network, and
the like for relay sections, the method being capable of notifying an upstream
or
a downstream client line of relay line fault information without causing link-
down
of a normal Ethemet path.
Another object of the present invention is to provide a wide area
Ethernet network which utilizes a plurality of transmission networks such as
the
GbE network, SONET/SDH network, and the like for relay sections, and which
is capable of notifying an upstream or a downstream client line of relay line
fault
information without causing link-down of a normal Ethernet path.
A further object of the present invention is to provide a wide area
Ethernet network which is capable of managing a transmission medium
network layer and a path network layer in a layered structure to provide
advanced network services.
According to an aspect of the present invention, there is provided an
Ethernet multiplexer for relaying information transmitted and received between
Ethernet networks for accommodating client lines from client terminals and a
plurality of types of transmission networks for connecting said Ethernet
11
CA 02519617 2005-09-14
networks to each other, the Ethemet multiplexer serving as an edge note of
each of the Ethernet networks, comprising a GbE_MAC processing circuit for
transmitting and receiving a frame to and from the transmission networks, and
for detecting a fault occurring on any of the transmission networks; a
TDM_MUX/DEMUX circuit for time-division-multiplexing or demultiplexing
frames transmitted in an upstream direction or in a downstream direction in
units of a plurality of Ethernet paths set between client terminals placed in
communication; an FE_MAC processing circuit for detecting transmission and
reception of an Ethernet frame to and from one of the client lines, and
detecting
a fault occurring on said one client line; and a GBP processing circuit, in
response to a received Ethernet frame sent from one of the client terminals,
for
dividing the Ethernet frame into a predetermined length to generate fixed-
length
frames, generating a capsule including a core block composed of each of the
fixed-length frames and a type field added thereto for notifying a fault
occurring
on said one client line, a forward relay line fault notification field for
notifying a
fault occurring in the transmission networks in a forward direction, and a
backward relay line fault notification field for notifying the fault in a
backward
direction, the GBP processing circuit, upon detection of a fault occurring on
one
of said client lines, adding a code indicative of the fault on said one client
line to
said type field, and overwriting a payload with a predefined idle frame, the
GBP
processing circuit, upon receipt of a frame from one of said transmission
networks, monitors the core block, the forward relay line fault notification
field,
and the backward relay line fault notification field, respectively, to
identify a fault
occurring on said one client line or a fault occurring in said transmission
networks for each of said Ethemet paths.
According to another aspect of the present invention, there is provided
an Ethernet multiplexer for relaying a capsule transmitted and received
12
CA 02519617 2005-09-14
between Ethernet networks each for accommodating client lines from client
terminals and a transmission network for interconnecting the Ethernet
networks, the capsule including a type field for notifying a fault occurring
on one
of said ciient lines, a forward relay line fault notification field for
notifying a fault
occurring in the transmission network in a forward direction, and a backward
relay line fault notification field for notifying the fault in a backward
direction, the
Ethernet multiplexer comprising a terminator circuit for transmitting and
receiving frames to and from the transmission network, and detecting a line
fault in the transmission network; a TDM_MUX/DEMUX circuit for
time-division-multiplexing or demultiplexing frames transmitted in an upstream
direction or in a downstream direction in a predefined order in units of a
plurality
of Ethernet paths which are set between client terminals placed in
communication; a MAC processing circuit for transmitting and receiving frames
to and from the Ethemet networks, and detecting a line fault in the Ethernet
networks; and GBP relay processing circuit for relaying the capsule for each
of
said Ethernet paths, and in response to a fault detected by the terminator
circuit
or the MAC processing circuit, setting a code indicative of the presence of a
line
fault in the forward relay line fault notification field, setting a code
indicative of
the absence of a line fault in the backward relay line fault notification
field,
setting a code indicative of no fault on said client lines in the type field,
and
setting a predefined idle frame in a payload.
According to the invention, which is claimed in parent Canadian Patent
Application Serial No. 2,458,694, there is provided an alarm transfer method
for
mutually notifying client terminals placed in communication that a fault
occurs in
a wide area Ethernet network which has Ethernet networks each for
accommodating client lines from the client terminals and a plurality of types
of
transmission networks for connecting the Ethernet networks to one another is
13
CA 02519617 2005-09-14
provided. The method includes the steps of dividing an Ethernet frame sent
from one of the client terminals into a predetermined fixed length to generate
a
plurality of fixed-length frames; generating capsules each comprising each of
the fixed-length frames, a type field for notifying a fault occurring on one
of the
client lines, a forward relay line fault notification field for notifying a
fault
occurring in the transmission network in a forward direction, and a backward
relay line fault notification field for notifying the fault in a backward
direction,
together with the fixed-length frame; multiplexing the generated capsules for
each of a plurality of Ethernet paths set between client terminals placed in
communication to generate a frame adapted for the transmission network, and
transferring the frame to an Ethernet network which accommodates a
destination client terminal; and demultiplexing the respective capsules from
the
received frame, and referencing the type field, the forward relay line fault
notification field, and the backward relay line fault notification field to
recognize
the fault occurring on the one client line or the fault occurring in the
transmission network for each of the Ethernet paths.
In this event, the forward relay line fault notification may be
transferred down to an egress node which is an edge node of an Ethernet
network that accommodates the destination client terminal, and a backward
relay line fault notification corresponding to the forward relay line fault
notification may be issued from the egress node to an ingress node which is an
edge node of an Ethernet network that accommodates a source client terminal.
Upon detection of the forward relay line fault notification, the egress
node may set a client line downstream of a corresponding Ethernet path to
link-down when the fault notification is not cleared even after the lapse of a
predetermined protection time. On the other hand, upon detection of the
14
CA 02519617 2005-09-14
backward relay line fault notification, the ingress node may set a client line
upstream of a corresponding Ethernet path to link-down.
According to the invention, which is claimed in parent Canadian Patent
Application Serial No. 2,458,694, there is provided a wide area Ethernet
network having Ethernet networks each for accommodating client lines from
client terminals, and a plurality of types of transmission networks for
interconnecting the Ethernet networks, wherein a fault occurring on one of the
client line or in the transmission network is mutually notified between the
client
terminals placed in communication. The wide area Ethernet network includes a
multiplexer which functions as an edge node of the Ethernet network. Upon
receipt of an Ethernet frame sent from one of the client terminals, the
multiplexer divides the Ethernet frame every predetermined fixed length to
generate a plurality of fixed-length frames, generates a capsule comprising
each of the fixed-length frames, a type field for notifying a fault occurring
on
one of the client lines, a forward relay line fault notification field for
notifying a
fault occurring in the transmission network in a forward direction, and a
backward relay line fault notification field for notifying the fault in a
backward
direction, multiplexes the generated capsules for each of a plurality of
Ethernet
paths set between client terminals placed in communication to generate a
frame adapted for the transmission network, and transfers the frame to an
Ethernet network which accommodates a destination client terminal. Upon
receipt of a frame from the transmission network, the multiplexer
demultiplexes
the respective capsules from the frame for each of Ethernet paths set between
client terminals placed in communication, and references the type field, the
forward relay line fault notification field, and the backward relay line fault
notification field to recognize the fault occurring on the one client line or
the
fault occurring in the transmission network for each of the Ethernet paths.
CA 02519617 2005-09-14
In this event, the multiplexer may issue the forward relay line fault
notification to an egress node which is an edge node of an Ethernet network
that accommodates a destination client terminal when the multiplexer operates
as an ingress node which is an edge node of the Ethernet network that
accommodates a source client terminal. Upon detection of the forward relay
line fault notification, the multiplexer may issue a backward relay line fault
notification corresponding to the forward relay line fault notification toward
the
ingress node, when the multiplexer operates as the egress node.
Alternatively, upon detection of the forward relay line fault notification,
the multiplexer may forcefully set a client line downstream of a corresponding
Ethemet path to link-down when the fault notification is not cleared even
after
the lapse of a predetermined protection time when the multiplexer operates as
the egress node. On the other hand, upon detection of the backward relay line
fault notification, the multiplexer may forcefully set a client line upstream
of a
corresponding Ethernet path immediately to link-down when the multiplexer
operates as the ingress node.
In the alarm transfer method and wide area Ethernet network as
described above, information on a fault occurring in a transmission network
can
be transferred both in the forward direction and in the backward direction by
providing the transport header of the GBP capsule with the forward relay line
fault notification field and backward relay line fault notification field.
Further, a
fault on a client line can be notified to a client terminal of a communication
partner using the type field in the GBP core block. Thus, according to the
invention, which is claimed in parent Canadian Patent Application Serial No.
2,458,694, the egress node can accomplish the link-pass-through for a
downstream client terminal in units of Ethernet paths, while the ingress node
16
CA 02519617 2005-09-14
can accomplish the link-pass-through for an upstream client terminal in units
of
Ethernet paths.
Also, by transferring the forward relay line fault notification down to the
egress node and issuing a backward relay line fault notification therefrom,
information on a fault occurring in a transmission network can be transferred
to
respective Ethemet networks which accommodate a source client terminal and
a destination client terminal, respectively, so that the link-pass-through of
a
relay line fault can be implemented even in a wide area Ethernet network which
includes a plurality of types of transmission networks. Further, an APS timer
circuit for measuring protection time may be provided only in the egress node.
Generally, since the APS timer circuit must be provided in each Ethernet path,
the circuit scale per node can be limited by providing the APS timer circuit
in
the single egress node than by providing the APS timer circuit in each of a
larger number of relay devices for multiplexing Ethernet paths.
According to the invention, which is claimed in parent Canadian Patent
Application Serial No. 2,458,694, there is provided a wide area Ethernet
network comprising of an ingress node for dividing a higher-level protocol
data
sent from each of a plurality of client terminals every predetermined fixed
length
to generate a plurality of fixed-length frames, generating a capsule
comprising
each of the fixed-length frames, a CRC field for detecting whether or not the
data sent from the each client terminal is normal, a type field for notifying
a fault
occurring on a client line, a forward relay line fault notification field for
notifying
a fault occurring in a transmission network in a forward direction, and a
backward relay line fault notification field for notifying the fault in a
backward
direction, multiplexing the capsules in a predefined order and adding an FCS
field for detecting whether or not data in the capsules is normal to generate
a
multiplexed MAC frame, and sending the multiplexed MAC frame; and an
17
CA 02519617 2005-09-14
egress node for detecting a defective data reception for each Ethernet path
from the result of checking the CRC field added to the fixed-length frame,
detecting a relay line fault in the forward direction and in the backward
direction
for each Ethernet path from information in the type field, the forward relay
line
fault notification field, and the backward relay line fault notification
field,
identifying an alarm for a path network layer from the detected defective data
reception and the relay line fault information, detecting a defective data
reception for each line from the result of checking the FCS field added to the
multiplexed MAC frame, and detecting loss of signal and link-down for each
line
to identify an alarm for a transmission medium network layer.
The wide area Ethemet network may further comprise a relay node for
relaying the multiplexed MAC frame transmitted/received between the ingress
node and the egress node by multiplexing frames defined by SONET/SDH into
the multiplexed MAC frame or demultiplexing the multiplexed MAC frame into
the SONET/SDH defined frames, identifying an alarm on a path-by-path basis
from a POH byte defined by the SONET/SDH, and identifying an alarm on a
section-by-section basis from an SOH byte defined by the SONET/SDH.
In the wide area Ethernet networks as described above, the transmission
medium network layer is monitored for a fault based on the result of a CRC
operation using the FCS field in the multiplexed MAC frame, and the path
network layer is monitored for a fault based on the result of a CRC operation
on
a CRC field in a GBP capsule of each Ethernet path, so that the section and
path can be managed independently of each other as in the SONET/SDH
network, and therefore, a transmission network can be organized in a layered
structure comprised of a transmission medium network layer and a path
network layer. Consequently, according to the invention, which is claimed in
parent Canadian Patent Application Serial No. 2,458,694, even in a wide area
18
CA 02519617 2005-09-14
Ethernet network which applies GBP encapsulation, in which an Ethernet
network is used as a transmission network for a relay section, the design,
maintenance, and operation of the network can be managed in a layered
structure. Thus providing advanced network services, as is the case with a
wide area Ethernet network which employs a SONET/SDH network in a relay
section.
In a wide area Ethemet network in which a transmission network in a
relay section is implemented by a combination of an Ethernet network and a
SONET/SDH network such that the SONET/SDH network is utilized as a
transmission network in a certain section and the GbE network is utilized as a
transmission network in another section transmission medium network layers
corresponding to respective sections can be managed irrespective of the type
of the transmission networks. In addition, the GBP capsule of the wide area
Ethernet network can be used to manage a path network layer from an ingress
node to an egress node on an end-to-end basis. Thus, according to the
invention, which is claimed in parent Canadian Patent Application Serial No.
2,458,694, even a wide area Ethernet network, which includes a combination of
a plurality of transmission networks, can be designed, operated, and
maintained without awareness of the difference between the transmission
networks, thereby providing advanced network services. Furthermore, in such
a configuration, it is possible to reduce the circuit scale and mounting area
of a
relay node which corresponds to a connection of a SONET/SDH network with
an Ethernet network. This is because the respective Ethernet networks are
provided with the same switching means as that used in the SONET/SDH
network to share the APS (automatic protection switching) processing circuit
provided for the SONET/SDH network. When a single APS processing circuit
19
CA 02519617 2005-09-14
can be shared for the SONET/SDH and Ethernet networks, the relay node can
be reduced in circuit scale and mounting area.
Furthermore, in the invention, which is claimed in parent Canadian
Patent Application Serial No. 2,458,694, APS processing can be performed as
fast as in the SONET/SDH network by using K1 byte and K2 byte and applying
similar APS processing to that used in the SONET/SDH network in a section in
which the Ethernet is used as a transmission network.
The above identified objects, features, and advantages of the present
invention will become apparent from the following description with reference
to
the accompanying drawings which illustrate examples of the present invention.
BRIEF DESCRIPTION OF THE DRAWINGS
Fig. 1 is a diagram showing a format for a prior art client management
frame defined in the GFP;
Fig. 2 is a block diagram illustrating an exemplary configuration of a
conventional expanded Ethemet network;
Figs. 3A to 3E are diagrams showing a format for a frame defined by
GBP (Generic Blocking Procedure);
Figs. 4A to 4C are schematic diagrams showing by way of example how
Ethernet frames for a number of lines m encapsulated by the GBP are
multiplexed on a single frame for a GbE line;
Figs. 5A to 5C are diagrams showing by way of example how n
multiplexed MAC frames shown in Figs. 4A to 4C are multiplexed on a single
frame for a SONET/SDH line;
Fig. 6 is a block diagram illustrating an exemplary configuration of a wide
area Ethernet network to which an alarm transfer method according to the
invention is applied;
CA 02519617 2005-09-14
Fig. 7 is a block diagram illustrating the configuration of an FE
multiplexer shown in Fig. 6;
Fig. 8 is a block diagram illustrating the configuration of a GbE
multiplexer shown in Fig. 6;
Fig. 9 is a block diagram illustrating an alarm transfer operation when a
fault occurs on a client line in the wide area Ethernet network illustrated in
Fig.
6;
Fig. 10 is a block diagram iilustrating an example of the alarm transfer
operation when a fault occurs on a relay GbE line in the wide area Ethernet
network illustrated in Fig. 6;
Fig. 11 is a block diagram illustrating a second example of the alarm
transfer operation when a fault occurs on the relay GbE line in the wide area
Ethernet network illustrated in Fig. 6;
Fig. 12 is a block diagram illustrating the alarm transfer operation when a
fault occurs on a relay 10G SONET/SDH line in the wide area Ethernet network
illustrated in Fig. 6;
Fig. 13 is a block diagram illustrating the configuration of a wide area
Ethernet network according to a second embodiment of the invention;
Fig. 14 is a block diagram illustrating the configuration of an FE
multiplexer shown in Fig. 13;
Fig. 15 is a table showing exemplary classifications for operations of an
alarm processing circuit contained in the FE multiplexer illustrated in Fig.
14;
Fig. 16 is a block diagram illustrating the configuration of a wide area
Ethernet network according to a third embodiment of the invention;
Fig. 17 is a block diagram illustrating the configuration of a GbE
multiplexer shown in Fig. 16; and
21
CA 02519617 2007-04-05
Fig. 18 is a table showing exemplary classifications for operations of an
alarm processing circuit contained in the GbE multiplexer illustrated in Fig.
17.
DETAILED DESCRIPTION OF THE INVENTION
First Embodiment:
The present invention employs GBP for protocol conversion between the
Ethernet and SONET/SDH. Details of the GBP is described in Japanese
Patent Application No. 2002-374791 (JP, P2002-374791) filed on Dec. 25,
2002 and United States Patent Application Publication No. 2004/0136401
published on July 15, 2004 correspond to Canadian Patent Application Serial
No. 2,453,738 filed on Dec. 19, 2003 and owned by the assignee of this
application.
Figs. 3A to 3E show a format for a frame defined by the GBP. According
to the GBP, an Ethernet frame as shown in Fig. 3A is divided into fixed length
payloads each having 16 octets from the top (see Fig. 3B). The Ethernet frame
has a variable length which is not always an integer multiple of 16 octets.
For
this reason, the last fixed length payload may include an extra pad region
comprised of a predetermined consistent data pattern, if the payload has less
than 16 octets, for filling the remaining free space. Subsequently, a header
field and the like is added to each of the fixed length payloads to generate a
GBP capsule of 18 octets long, as shown in Fig. 3C.
The structure of the GBP capsule is shown in Figs. 3D and 3E.
Specifically, Fig. 3D shows a frame format in normal operation, and Fig. 3E
shows a frame format when a fault occurs on a relay line. As shown in Figs. 3D
and 3E, the GBP capsule has a GBP transport header, a GBP core block, and
an 8-bit CRC (Cyclic Redundancy Check) field. The CRC operation is
performed for both the GBP transport header and GBP core block.
22
CA 02519617 2005-09-14
The GBP core block has a 5-bit type field in front of the fixed length
payload. The type field is defined for control, such that a value in this
field is
relied on to notify a fault on a client line, as is the case with the
aforementioned
GFP.
The GBP transport header is a 3-bit field comprised of a 2-bit undefined
area and a 1-bit GBP core block indicator field. However, as shown in Fig. 3E,
when a fault occurs on a relay line, the undefined area is replaced with a 1-
bit
forward relay line fault notification field and a 1-bit backward relay line
fault
notification field.
The GBP core block indicator field stores a value indicating whether the
subsequent GBP core block is valid or invalid, where "1" indicates that the
GBP
core block is valid, and "0" indicates that the GBP core block is invalid
(idle).
The core block indicator field is set to "0" in response to detection of the
forward relay line fault notification, which will be described later.
The forward relay line fault notification field is used to notify the
presence or absence of a fault on a relay line in the forward direction, while
the
backward relay line fault notification field is used to notify the presence or
absence of a fault on a relay line in the backward direction.
Figs. 4A to 4C show, by way of example, how Ethernet paths (client
lines) having m number of lines (m is an arbitrary positive integer),
encapsulated by GBP are multiplexed on a single relay GbE line. The
multiplexed client lines are assumed to be m fast Ethernet (hereinafter
abbreviated as "FE") lines.
As shown in Figs. 4A and 413, GBP capsules for m lines are sequentially
time-division-multiplexed on a relay GbE line, and the time-division-
multiplexing
is repeated k times (k is an arbitrary positive integer) to generate one
frame. In
this example, k is a value determined by the relationship between the
23
CA 02519617 2005-09-14
transmission rate on the GbE line and the transmission rate on the client
line.
While Fig. 4A only describes an Ethernet frame transferred on a single client
line (Ethernet path #1), Ethernet frames are actually transferred on m client
lines (Ethernet paths #1-#m), so that GBP capsules are extracted in order from
the first one in each of the Ethernet frames for time division multiplexing on
a
relay GbE line.
Also, as shown in Fig. 4C, a multiplexed MAC frame header is added in
front of the first GBP capsule after the multiplexing. The multiplexed MAC
frame header is comprised of a sequence number storing field (of two octets
long); a K1 byte storing field (of one octet long); a K2 byte storing field
(of one
octet long); a reserved field (of eight octets long); and an HEC (Header Error
Control) field (of two octets long).
Further, a header (MAC frame header) including a destination address,
source address, priority information, and the like is added in front of the
multiplexed MAC frame header, and an FCS field is added to the end of the
frame. The resulting frame is called the "multiplexed MAC frame. "
Figs. 5A to 5C show by way of example how n multiplexed MAC frames
shown in Figs. 4A to 4C are multiplexed on a single frame for a SONET/SDH
line. Specifically, Figs. 5A to 5C show an example in which n multiplexed MAC
frames are separated into mxn GBP capsules, and the separated mxn Ethernet
frames (i. e., GBP capsules) are multiplexed on a single SONET/SDH line.
As shown in Figs. 5A to 5C, in the SONET/SDH network, one Ethernet
path on an FE line is assigned to one VC-4 (155. 52 Mbps) frame, which is a
path defined by SONET/SDH, and the VC-4 frame stores data on the Ethernet
path in its payioad.
On the other hand, when a ciient line is a GbE line, an Ethernet path on
one GbE line is assigned to VC-4 frames for eight channels, and the VC-4
24
CA 02519617 2005-09-14
frames for eight channels store data on the Ethernet path in their payloads.
One GbE line can be stored in payloads of a plurality of channels, for
example,
by applying the virtual concatenation described on page 108 of ITU-T
Recommendation G. 707 (October 2000). While the foregoing example
provides the VC-4 frames for eight channels, this is only an example, and the
VC-4 frames for any number of channels may be provided as long as they are
equal to or more than the bandwidth of the GBP encapsulated GbE line.
Alternatively, a plurality of VC-3 frames may be used.
Virtual concatenation may also be used in the aforementioned transfer of
the Ethernet path on the FE line, in which case an Ethernet path on one FE
line
may be assigned to VC-3 frames for a plurality of channels, and the VC-3
frames for the plurality of channels may store data on the Ethemet path in
their
payloads.
The Ethernet frame (GBP capsule. See Fig. 5A) of 18 octets long
transferred along an Ethernet path is delimited every octet, and stored in a
payload of an OC-192 frame defined by SONET/SDH in units of VC-4 (for each
Ethernet path), as shown in Fig. 5C. In this event, idle blocks are inserted
and
extracted, compensating for the difference between the line rate of the VC-4
frames and the line rate of the GBP capsules.
A section overhead of the OC-1 92 frame shown in Fig. 5C stores
network management information such as frame synchronization, error
monitoring, alarm transfer, and the like, and an AU pointer stores information
for indicating the head position of a VC-4 frame. The VC-4 frame has a POH
(Path Over Head) field which stores an error monitoring alarm. These fields
are
all defined by SONET/SDH.
Fig. 6 illustrates the configuration of a wide area Ethernet network to
which the aiarm transfer method is applied in accordance with the first
CA 02519617 2005-09-14
embodiment. Fig. 7 illustrates the configuration of an FE multiplexer shown in
Fig. 6, and Fig. 8 illustrates the configuration of a GbE multiplexer shown in
Fig.
6.
In the first embodiment, an alarm is transferred upon detection of a fault
using the type field in the GBP core block for transferring a fault occurring
on a
client line. In addition, the alarm transfer employs a forward relay line
fault
notification field and backward relay line fault notification field provided
in the
GBP transport header. Upon detection of a fault on an upstream line at a relay
node, a forward relay line fault notification is issued to all Ethernet paths
that
pass the line on which the fault has occurred. Also, from an egress node,
which has received the forward relay line fault notification, to an upstream
ingress node, a backward relay line fault notification is issued only to those
Ethernet paths which have received the forward relay line fault notification.
In
this example, if the forward relay line fault notification is still received
even after
the lapse of an arbitrary protection time, set by an APS timer circuit, the
backward relay line fault notification is issued. Upon receipt of the backward
relay line fault notification, the ingress node forcefully brings the client
line
associated with the Ethernet path into link-down.
As illustrated in Fig. 6, the wide area Ethernet network in the first
embodiment comprises a plurality of Ethernet networks (two are shown in Fig.
6), each of which accommodate a plurality of client terminals, that are
connected through GbE networks and a SONET/SDH network. The Ethernet
networks used herein may be FE networks, which provide the transmission
capability of 10 Mbps or 100 Mbps. The SONET/SDH network used herein
may be 10G SONET/SDH network which has the transmission capability of 9.
953 Gbps per port.
26
CA 02519617 2005-09-14
Client terminals 1-1-1 to 1-n-m accommodated in one FE network shown
in Fig. 6 are connected to FE multiplexers 2-1 to 2-n, which are edge nodes,
respectively, while client terminals 7-1-1 to 7-n-m accommodated in the other
FE network are connected to FE multiplexers 6-1 to 6-n, which are edge nodes,
respectively. Client terminals 1-1-1 to 1-n-m are identical in configuration
to
client terminals 7-1-1 to 7-n-m, though they are installed at locations
opposite to
each other across the relay lines. Each of the client terminals may be an
Ethernet switch such as a hub.
GbE multiplexer 3 is a relay device in a GbE network which
accommodates lines directed to FE multiplexers 2-1 to 2-n, while GbE
multiplexer 5 is a relay device in a GbE network which accommodates lines
directed to FE multiplexers 6-1 to 6-n. GbE multiplexer 3 is identical in
configuration to GbE multiplexer 5, though they are installed at locations
opposite to each other across the relay line. SONET/SDH cross-connect (XC)
device 4 is a relay device in the SONET/SDH network for relaying GbE
multiplexer 3 to GbE multiplexer 5 and vice versa.
In the following description, each line or circuit which connects a client
terminal to an FE multiplexer shown in Fig. 6 is called the "client line," and
a
line or circuit which connects an FE multiplexer to a GbE multiplexer is
called
the "relay GbE line. " Also, a line or circuit which connects a GbE
multiplexer to
the SONET/SDH cross-connect device is called the "relay SONET/SDH line. "
Here, the client line section is operated with an auto-negotiation
(AUTONEG) function, defined by the Ethernet, being set enabled. The auto
negotiation function may be set disabled in a certain network configuration.
The relay GbE line section is operated with the auto-negotiation function
being
set disabled.
27
CA 02519617 2005-09-14
The auto-negotiation function involves mutually communicating
information between devices interconnected through an Ethernet line to set an
optimal communication mode (transmission rate, full duplex/half duplex), and
communicating line fault information to set a backward line to link-down when
a
fault occurs on a forward line. Information for the auto-negotiation function
may
be communicated by an FLP (Fast Link Pulse) on the FE network.
SONET/SDH cross-connect device 4 is provided for switching
transmission paths of frames, used in the 10G SONET/SDH network.
FE multiplexers 2-1 to 2-n, and 6-1 to 6-n each time-division-multiplex m
FE frames on a single GbE frame, and also demultiplex one GbE frame into m
FE frames. For example, FE multiplexer 2-1 transforms FE frames transmitted
from client terminals 1-1-1 to 1-1-m into GBP capsules, and
time-division-multiplexes the GBP capsules for one GbE line to generate a
multiplexed MAC frame. FE multiplexer 2-1 also demultiplexes a GbE frame
received from one GbE line into m FE frames.
GbE multiplexer 3 accommodates GbE lines from FE multiplexers 2-1 to
2-n, and time-division-multiplexes frames received from n GbE lines on a
single
10G SONET/SDH line. GbE multiplexer 3 also demultiplexes a SONET/SDH
frame received from the 10G SONET/SDH line into n GbE frames.
The number m of multiplexing provided by respective FE multiplexers 2-1 to
2-n, and 6-1 to 6-n is set, for example, to a positive integer equal to or
less than
eight from the relationship between the capacity of the GbE line and an FE
frame transfer band. Also, the number n of multiplexing provided by the
respective GbE multiplexers is also set to a positive integer equal to or less
than eight for the same reason.
For implementing communication between client terminals opposing
each other across the relay lines in the wide area Ethernet network
illustrated in
28
CA 02519617 2005-09-14
Fig. 6, a source client terminal and a destination client terminal are
determined,
multiplexing/demultiplexing orders in associated FE multiplexers and GbE
multiplexers in accordance with a transmission path which interconnects these
client terminals, and a switching path are all determined in the SONET/SDH
cross-connect device. By determining line settings in each node in accordance
with the source client terminal and destination client terminal, Ethernet
frames
sent from an arbitrary client terminal are transferred only to the determined
client terminal. A fixed transmission path built by the line settings for
transmitting Ethernet frames is called the "Ethemet path. " Here, such a
transmission path is called the "FE path" when a client line is an FE line,
and is
called the "GbE path" when a client line is a GbE path.
Fig. 6 illustrates an example in which FE paths are set between ciient
terminal 1-1-1 and client terminal 7-1 -1, wherein FE path 8 is an Ethernet
path
which defines the forward direction from client terminal 1-1-1 to client
terminal
7-1-1, and FE path 9 is an Ethernet path in the backward direction with
respect
to FE path 8, and is also an Ethernet path which defines the forward direction
from client terminal 7-1-1 to client terminal 1-1-1.
As illustrated in Fig. 7, FE multiplexers 2-1 to 2-n and 6-1 to 6-n each
comprise GbE MAC processing unit 19 for performing MAC processing such as
transmission/reception of frames, detection of a line fault, and the like for
the
associated relay GbE line; TDM MUX/DEMUX circuit 18 for
time-division-multiplexing. or demultiplexing frames transmitted in the
forward
direction or in the backward direction in a predefined order; GBP processing
circuits 17-1 to 17-m each for generating and terminating GBP capsules for
each FE path and having an alarm transfer function; and MAC processing
circuits 16-1 to 16-m each for performing MAC processing such as
29
CA 02519617 2005-09-14
transmission/reception of frames, detection of a line fault, and the like for
the
associated client line.
GbE MAC processing unit 19 comprises O/E circuit 36 for converting a
received optical signal to an electric signal; GbE MAC termination processing
circuit 37 for detecting reception of an FE frame, and a fault on a relay GbE
line; E/O circuit 38 for converting an electric signal to an optical signal
which is
sent out therefrom; and GbE MAC generation processing circuit 39 for
generating GbE frames.
TDM MUX/DEMUX circuit 18 comprises TDM DEMUX processing circuit
40 for demultiplexing a received frame in accordance with a predefined order;
and TDM MUX processing circuit 41 for time-division-multiplexing frames in a
predefined order.
GBP processing circuits 17-1 to 17-m each comprise GBP termination
processing circuit 42; holding circuit 51 for holding a client line fault
notification,
a forward relay line fault notification, and a backward relay line fault
notification
delivered from GBP termination processing circuit 42 for a fixed time period;
OR circuit 43 for taking logical OR of the forward relay line fault
notification and
a GbE line fault detection signal delivered from GbE MAC processing unit 19;
APS timer circuit 45 triggered by an output signal of OR circuit 43 to start
counting and count up to TAPS [ms] (approximately 50 milliseconds) while the
output signal remains at "1 "; holding circuit 47 for holding the output
signal of
OR circuit 43 at "1 "(signal for instructing issuance of a backward relay line
fault
notification) until the output signal of OR circuit 43 transitions to "0"
after APS
timer circuit 45 has counted up to TAPS [ms]; OR circuit 44 for taking logical
OR
of the output signal of holding circuit 47 and the client line fault
notification and
backward relay line fault notification delivered from holding circuit 51; and
GBP
CA 02519617 2005-09-14
generation processing circuit 48 for generating GBP capsuies and an alarm for
a client line fault.
APS timer circuit 45 measures the protection time required for switching
a transmission path in order to provide redundancy for a relay line which
interconnects client terminals. If the forward relay line fault notification
is still
being received even after the lapse of time TAPS defined by APS timer circuit
45, a transmission path of a relay line is switched.
FE MAC processing circuits 16-1 to 16-n each comprise FE MAC
generation processing circuit 49 for generating FE frames; physical device 93
for transmitting FE frames; physical device 94 for receiving FE frames from an
associated FE line; and FE MAC termination processing circuit 50 for receiving
an FE frame and detecting a client line fault.
As illustrated in Fig. 8, GbE multiplexers 3, 5 each comprise 10G
SONET generator/terminator circuit 23 for frames to/from the SONET/SDH line
and detecting a line fault on the SONET/SDH line; TDM MUX/DEMUX circuit 22
for time-division-multiplexing or demultiplexing frames to be transmitted in
the
upstream direction or in the downstream direction in a predefined order; GBP
relay processing circuits 21-1 to 21-n each for performing operations
associated with relay processing such as a rate adjustment of the GBP capsule
for each GbE path, and having an alarm transfer function; and GbE MAC
processing units 20-1 to 20-n each for performing MAC processing such as
transmission/reception of frames to/from a GbE line, detection of a line
fault,
and the like.
10G SONET generator/terminator circuit 23 comprises an O/E circuit 52
for converting a received optical signal to an electric signal; SONET/SDH
reception processing circuit 53 for detecting reception of a SONET/SDH frame
and a fault on the SONET/SDH line; E/O circuit 54 for converting an electric
31
CA 02519617 2005-09-14
signal to an optical signal which is sent out therefrom; and SONET/SDH
transmission processing circuit 55 for generating SONET/SDH frames.
TDM MUX/DEMUX circuit 22 comprises TDM DEMUX processing circuit
56 for demultiplexing a received frame in accordance with a predefined order;
and TDM MUX processing circuit 57 for time-division-multiplexing data in a
predefined order.
GBP relay processing circuits 21-1 to 21-n each comprise P-AIS detector
circuit 62 for detecting path AIS in a VC-4 frame for each FE path; first GBP
relay processing circuit 58 for relaying from VC-4 to GbE; and second GBP
relay processing circuit 59 for relaying from GbE to VC-4.
GbE MAC processing units 20-1 to 20-n each comprise GbE MAC
generation processing circuit 60 for generating GbE frames; E/O circuit 95 for
transmitting generated GbE frames; O/E circuit 96 for receiving frames from a
GbE line; and GbE MAC termination processing circuit 61 for receiving a GbE
frame and detecting a fault on a GbE line.
It should be noted that "MAC" within FE multiplexers 2-1 to 2-n and 6-1
to 6-n shown in Fig. 6 corresponds to FE MAC processing circuits 16-1 to 16-n
shown in Fig. 7, and "GBP" corresponds to GBP processing circuits 17-i to
17-n shown in Fig. 7. Also, "MUX/DEMUX" corresponds to TDM MUX/DEMUX
circuit 18 shown in Fig. 7, and "GbE MAC" corresponds to GbE MAC
processing unit 19 shown in Fig. 7.
"GbE MAC" within GbE multiplexers 3, 5 shown in Fig. 6 corresponds to
GbE MAC processing units 20-1 to 20-n shown in Fig. 8, "GBP R" corresponds
to GBP relay processing circuits 21-1 to 21-n shown in Fig. 8. Also,
"MUX/DEMUX" corresponds to TDM MUX/DEMUX circuit 22 shown in Fig. 8,
and "SONET" corresponds to 10G SONET generator/terminator circuit 23
shown in Fig. 8.
32
CA 02519617 2005-09-14
Description will be next made on the operation of the FE multiplexers
and GbE multiplexers during an alarm transfer.
First, the operation of FE multiplexers 2-1 to 2-n and 6-1 to 6-n will be
described with reference to Fig. 7.
In each of FE multiplexers 2-1 to 2-n and 6-1 to 6-n, O/E circuit 36
converts main signal data (GbE frames) received from a relay GbE line to an
electric signal, and GbE MAC termination processing circuit 37 performs MAC
termination processing such as removal of a preamble, FCS checking, and the
like.
The MAC termination processing also includes detection of link-down.
Upon detection of link-down, MAC termination processing circuit 37 applies all
GBP processing circuits 17-1 to 17-m with a GbE line fault detection signal at
"1. " The GbE line fault detection signal is continuously delivered while
link-down is being detected. MAC termination processing circuit 37 delivers
the
GbE line fault detection signal at "0" at the time of link up.
The frames delivered from GbE MAC termination processing circuit 37 is
the aforementioned multiplexed MAC frame which is applied to TDM DEMUX
processing circuit 40.
TDM DEMUX processing circuit 40 demultiplexes the frame for each FE
path in accordance with a predefined order, and applies data demultiplexed
from the frame to GBP processing circuits 17-1 to 17-m which are associated
with defined FE ports (#1-#m).
Now, description will be made on the operation of GBP processing
circuits 17-1 to 17-m.
A frame applied to GBP termination processing circuit 42 in each of GBP
processing circuits 17-1 to 17-m is a FE frame corresponding to one FE path.
GBP termination processing circuit 42 terminates the frame to extract a GBP
33
CA 02519617 2005-09-14
capsule. Then, GBP termination processing circuit 42 examines the type field
in the GBP core block and the contents of the GBP transport header in the
extracted GBP capsule to detect a client line fault notification, a forward
relay
line fault notification, or a backward relay line fault notification which is
delivered to holding circuit 51. Holding circuit 51 holds the value of each
line
fault notification until the next GBP transport header is received.
OR circuit 43 takes iogical OR of the held forward relay line fault
notification, and a GbE line fault detection signal delivered from GbE MAC
termination processing circuit 37, and delivers the result of the logical OR
operation. When OR circuit 43 delivers the value of "1," this indicates that a
relay line fault has occurred on a FE path on which the forward direction is
defined from a relay GbE line to a client line.
APS timer circuit 45 is triggered by a transition of the output value of the
OR circuit 43 from "0" to "1 " to start counting up. Then, APS timer circuit
45
continues to count up when OR circuit 43 applies the value of "1 " to APS
timer
circuit 45, and applies a signal "1" to holding circuit 47 after it has
counted up to
maximum TAPS [ms]. APS timer circuit 45 resets the counted value to an initial
value "0" when OR circuit 43 applies the value of "0" to APS timer circuit 45.
Holding circuit 47 holds the output at "1" until the output of OR circuit 43
transitions to "0" when APS timer circuit 45 delivers the value of I. "
OR circuit 44 takes logical OR of the output signal of holding circuit 47
and the client line fault notification and backward relay line fault
notification
delivered from holding circuit 51, and sends the result of the OR operation to
FE MAC generation processing circuit 49 as an FE line forcible shut-down
signal.
FE MAC generation processing circuit 49 performs MAC processing on
an FE frame delivered from GBP termination processing circuit 42 such as
34
CA 02519617 2005-09-14
addition of a preamble, and the like. The FE frame thus generated is sent onto
a client line from physical device 93.
Upon detection of the FE line forcible shut-down signal at "1" delivered
from OR circuit 44, FE MAC generation processing circuit 49 stops delivering
main signal data, and forcefully brings the client line into link-down.
In this way, upon detection of a client line fault notification or a backward
relay line fault notification, a downstream client line is immediately set
into
link-down in a forcible manner. Upon detection of a forward relay line fault
notification, on the other hand, a client line downstream of the FE path is
forcefully set into link-down after the lapse of protection time TAPs set in
APS
timer 45 to accomplish the link-pass-through for the downstream client line.
On the other hand, each of FE multiplexers 2-1 to 2-n and 6-1 to 6-n,
receives main signal data (FE frame) at physical device 94 from an associated
client line, FE MAC termination processing circuit 50 performs MAC termination
processing on the received data such as removal of the preamble, FCS
checking, and the like. FE MAC termination processing circuit 50 also detects
link-down of a client line, and upon detection of link-down, continuously
delivers
an FE line fault detection signal at "1 " to GBP generation processing circuit
48.
GBP generation processing circuit 48 generates a GBP capsule of 18
octets long from the FE frame applied thereto. In this event, GBP generation
processing circuit 48 references the FE line fault detection signal and the
backward relay line fault notification issuance instruction signal delivered
from
holding circuit 47, and sets a code ("0") indicative of the absence of a line
fault
in the backward rely line fault notification field in the GBP capsule, and
sets a
code indicative of the absence of a line fault in the type field in the GBP
capsule when the FE line fault detection signal is at "0" and the backward
relay
line fault notification issuance instruction signal is at "0. " In this event,
the
CA 02519617 2005-09-14
fixed-length payload field in the GBP core block is not overwritten, and is
passed through GBP generation processing circuit 48 without modification.
When the FE line fault detection signal is at " 1 " and the backward relay
line fault notification issuance instruction signal is at "0," GBP generation
processing circuit 48 sets the code ("0") indicative of the absence of a line
fault
in the backward relay line fault notification field in the GBP capsule, and
sets a
code indicative of the presence of a client line fault in the type field in
the GBP
capsule. In this event, the fixed-length payload field in the GBP core block
is
overwritten with a predefined fixed pattern (a pattern entirely composed of "1
or the like).
When the FE line fault detection signal is at "0" and the backward relay
line fault notification issuance instruction signal is at "1," GBP generation
processing circuit 48 sets a code ("0") indicative of the absence of a line
fault in
the forward relay line fault notification field in the GBP capsule, sets a
code
(1 ") indicative of the presence of a line fault in the backward relay line
fault
notification field, and sets a code indicative of no client line fault in the
type
field. In this event, the fixed-length payload field in the GBP core block is
not
overwritten, and is passed through GBP generation processing circuit 48
without modification.
When the FE line fault detection signal is at "1" and the backward relay
line fault notification issuance instruction signal is at "1," GBP generation
processing circuit 48 sets the code ("0") indicative of the absence of a line
fault
in the forward relay line fault notification field in the GBP capsule, sets
the code
("1 ") indicative of the presence of a line fault in the backward relay line
fault
notification field, and sets the code indicative of no client line fault in
the type
field. In this event, the fixed-length payload field in the GBP core block is
36
CA 02519617 2005-09-14
overwritten with a predefined fixed pattern (a pattern entirely composed of
"1"
or the like).
The GBP capsule thus generated is applied to TDM MUX processing
circuit 41 which time-division-multiplexes the GBP capsule with GBP capsules
of other FE paths, as shown in Fig. 4. Then, GbE MAC generation processing
circuit 39 adds a preamble signal and FCS to the multiplexed GBP capsules,
and delivers the resulting multiplexed MAC frame onto a relay GbE line through
E/O circuit 38.
Next, the operation of GbE multiplexers 3, 5 will be described with
reference to Fig. 8.
As illustrated in Fig. 8, in each of GbE multiplexers 3, 5, O/E circuit 52
converts main signal data (SONET/SDH frame) received from a 10G
SONET/SDH line to an electric signal, and SONET/SDH reception processing
circuit 53 executes the termination processing on the section overhead (SOH)
and path overhead (POH). In this event, upon detection of a path alarm
defined by SONET/SDH, SONET/SDH reception processing circuit 53 issues a
path AIS alarm to a corresponding VC-4 frame, and overwrites its payload field
entirely with "1. " Upon detection of a section alarm, SONET/SDH reception
processing circuit 53 issues a path AIS alarm to all VC-4 frames which pass
through the lOG SONET/SDH line, and overwrites their payload fields entirely
with "1. " The path AIS alarm is continuously issued while the path alarm or
section alarm is being detected.
A frame delivered from SONET/SDH reception processing circuit 53 is a
VC-4*64 (OC-192) which is applied to TDM DEMUX processing circuit 56.
TDM DEMUX processing circuit 56 demultiplexes an OC-192 frame in units of
VC-4 (FE path) in accordance with a predefined order, and delivers the
37
CA 02519617 2005-09-14
resulting data to associated GBP relay processing circuits 21-2 to 21-n,
respectively, for corresponding GbE paths.
Next, description will be made on the operation of GBP relay processing
circuits 21-1 to 21-n.
In each of GBP relay processing circuits 21-1 to 21-n, P-AIS detector
circuit 62 checks whether or not the payload field in a received frame is
fully
filled with "1" in units of VC-4, and determines that a path AIS alarm has
been
issued for the VC-4 when the payload field is fully filled with "1," and
delivers a
forward relay line fault notification issuance instruction signal at "1. " On
the
other hand, when the payload field is not fully filled with "1," P-AIS
detector
circuit 62 delivers the forward relay line fault notification issuance
instruction
signal at T. "
First GBP relay processing circuit 58 extracts GBP capsules from the
main signal data (VC-4xm) received thereby, and applies the rate on the GbE
line to the extracted GBP capsules. In this event, for adjusting the
difference in
rate between the m VC-4 frames and the GbE line, first GBP relay processing
circuit 58 thins out idle frames from the VC-4 frames.
When the forward relay line fault notification issuance instruction signal
is at "1," first GBP relay processing circuit 58 sets a code ("1 ") indicative
of the
presence of a line fault in the forward relay line fault notification field in
the GBP
transport header of the associated FE path. While first GBP relay processing
circuit 58 allows the fixed-length payload field in the GBP core block to pass
therethrough as it is, the payload field has been entirely overwritten with
"1" by
SONET/SDH reception processing circuit 53, so that no data is transferred in
the downstream direction. In other words, the occurrence of a forward relay
line fault is detected on the FE path on which the forward direction is
defined
from the relay 10G SONET/SDH line to the relay GbE line.
38
CA 02519617 2005-09-14
When the forward relay line fault notification issuance instruction signal
is at "0," first GBP relay processing circuit 58 allows the forward relay line
fault
notification field to pass therethrough as it is. First GBP relay processing
circuit
58 also allows the backward relay line fault notification field, the type
field in the
GBP core block, and the fixed-length payload to pass therethrough without
modification.
The main signal data generated in the foregoing manner is applied to
GbE MAC generation processing circuit 60 which generates a multiplexed MAC
frame which has the main signal data stored in its payload. The generated
multiplexed MAC frame is sent from E/O circuit 95 onto a relay GbE line.
On the other hand, upon receipt of main signal data from a GbE line, the
main signal data is converted to an electric signal by O/E circuit 96, and
undergoes the MAC termination processing such as removal of the preamble,
FCS checking, and the like in GbE MAC termination processing circuit 61.
During this event, GbE MAC termination processing circuit 61 delivers a GbE
line fault detection signal at "1" in response to detection of link-down on a
relay
GbE line.
A signal applied to second GBP relay processing circuit 59 is a
multiplexed MAC frame which has m FE paths multiplexed thereon. Second
GBP relay processing circuit 59 extracts GBP capsules for each FE path for
storage in a VC-4 frame associated with each FE path. Then, second GBP
relay processing circuit 59 inserts idle frames as appropriate for adjusting
the
rate with the VC-4 frame.
When the GbE line fault detection signal received from GbE MAC
generation processing circuit 60 is at "1," second GBP relay processing
circuit
59 sets a code ("1 ") indicative of the presence of a line fault in the
forward relay
line fault notification field in the GBP transport header of every FE path
which
39
CA 02519617 2005-09-14
passes through second GBP relay processing circuit 59. Further, second GBP
relay processing circuit 59 overwrites the fixed-length payload in the GBP
core
block with a fixed pattem (entirely composed of "1" or the like) indicative of
a
predefined idle frame. In other words, the occurrence of a fault is detected
on a
relay line of the FE path on which the forward direction is defined from the
relay
GbE line to the relay 10G SONET/SDH line.
When the GbE line fault detection signal is at "0," second GBP relay
processing circuit 59 allows the forward relay line fault notification field
to pass
therethrough without modifiction. Also, second GBP relay processing circuit 59
allows the backward relay line fault notification field, the type field in the
GBP
core block, and the fixed-length payload to pass therethrough without
modification.
The frame thus generated is applied to TDM MUX processing circuit 51
which multiplexes this frame with FE paths applied from other GbE ports. They
are multiplexed in a predefined order. Then, SONET/SDH transmission
processing circuit 55 adds a section overhead (SOH) and a path overhead
(POH), and the resulting multiplexed frame is sent from E/O circuit 54 to a
relay
lOG SONET/SDH line.
Next, description will be made on the operation when a fault occurs on a
relay line in the wide area Ethernet network illustrated in Fig. 6. When no
fault
occurs on any of client lines and relay lines, GBP relay processing circuit 21-
1
and GBP processing circuit 17-1 each set a code indicative of the absence of a
line fault in the GBP transport header, the forward relay line fault
notification
field and the backward relay line fault notification field of the GBP capsule.
During this event, a code indicative of no client line fault is also set in
the type
field in the GBP core block.
CA 02519617 2005-09-14
First, an alarm transfer operation will be described when a fault occurs
on a client line in the situation as mentioned above.
Fig. 9 illustrates an alarm transfer operation when a fault occurs on a
client line in the wide area Ethernet network illustrated in Fig. 6. As
illustrated
in Fig. 9, a fault occurs on a client line which interconnects client terminal
1-1-1
and FE multiplexer 2-1, and as link-down is accordingly detected by FE MAC
processing circuit 16-1 in FE multiplexer 2-1, GBP processing circuit 17-1 in
FE
multiplexer 2-1 sets a code indicative of a client line fault notification in
the type
field in the GBP core block of a GBP capsule which is transferred through the
FE path on which the fault has occurred. In this event, no change is made to
the GBP transport header. The fixed-length payload in the GBP core block is
overwritten with a predefined code indicative of an idle frame (entirely
composed of 1 " or the like), so that no client data is transferred in the
downstream direction.
This GBP capsule is transferred down to GBP processing circuit 17-1 in
opposing FE multiplexer 6-1, so that GBP processing circuit 17-1 detects a
client fault notification in the GBP core block. FE multiplexer 6-1 forcefully
transitions only a client line downstream of FE MAC processing circuit 16-1 to
link-down to accompiish the link-pass-through only for the downstream of the
FE path on which the client line fault has occurred. During this event, the
alarm
is not transferred in the backward direction of that path or to other FE
paths.
Next, description will be given on an alarm transfer operation when a
fault occurs on a relay GbE line.
Fig. 10 illustrates an alarm transfer operation when a fault occurs on a
relay GbE line in the wide area Ethernet network illustrated in Fig. 6. Since
the
relay GbE line is operated with the auto-negotiation function being disabled
as
41
CA 02519617 2005-09-14
mentioned above, a line fault in the forward direction, if any, would not
cause a
transition to link-down of the same section in the backward direction.
As illustrated in Fig. 10, a fault occurs on a relay GbE line which
interconnects FE multiplexer 2-1 and GbE multiplexer 3, and as link-down is
accordingly detected by GbE MAC processing unit 20-1 in GbE multiplexer 3,
GBP relay processing circuit 21-1 of GbE multiplexer 3 sets a code indicative
of
the presence of a line fault in the forward relay line fault notification
field in the
GBP transport header of a corresponding GBP capsule, sets a code ("0")
indicative of the absence of a line fault in the backward relay line fault
notification field, and sets a code indicative of no client line fault in the
type field
in the GBP core block. GBP relay processing circuit 21-1 also sets the fixed
payload to a predefined pattern (entirely composed of "1 " or the like)
indicative
of an idle frame. This processing is performed on every FE path which passes
along the relay GbE line on which the line fault has occurred.
The frame generated by GBP relay processing circuit 21-1 in the GbE
multiplexer 3 is transferred down to GBP processing circuit 17-1 in opposing
FE
multiplexer 6-1, where GBP processing circuit 17-1 detects the forward relay
line fault notification in the GBP transport header. When the forward relay
line
fault notification is not cleared after the lapse of APS protection time TAPS,
FE
multiplexer 6-1 forcefully transitions only a client line downstream of FE MAC
processing circuit 16-1 to link-down to accomplish the link-pass-through for
the
downstream. This I ink-pass-th rough processing is performed on every Ethernet
path which passes along the relay GbE line on which the line fault has
occurred. Then, GBP processing circuit 17-i sets a code ("0") indicative of
the
absence of a line fault in the forward relay line fault notification field in
the GBP
transport header of a corresponding GBP capsule, sets a code ("1 ") indicative
of the presence of a line fault in the backward relay line fault notification
field,
42
CA 02519617 2005-09-14
sets a code indicative of no client line fault in the type field in the GBP
core
block, and allows the fixed-length payload to pass therethrough as it is. The
resulting multiplexed MAC frame is sent in the backward direction. This
processing is performed on every FE path which imposes the forced down-link
to downstream client lines.
The frame delivered from GBP processing circuit 17-1 is transferred
down to opposing FE multiplexer 2-1 which detects the backward relay line
fault
notification in the GBP transport header. In this event, FE multiplexer 2-1
forcefully transitions only a client line downstream of FE MAC processing
circuit
16-1 to link-down.
As described above, since the auto-negotiation function is enabled in the
client line section, FE paths in the forward direction also transition to link-
down.
In this way, an alarm is transferred to upstream client terminal 1-1-1 to
accomplish the link-pass-through for the upstream client terminal. This
processing is performed on every FE path which has undergone the
iink-pass-through processing in the downstream direction resulting from the
fault on the relay GbE line.
Next, description will be made on an second example of the alarm
transfer operation when a fault occurs on a relay GbE line.
Fig. 11 illustrates a second alarm transfer operation when a fault occurs
on a relay GbE line in the wide area Ethernet network illustrated in Fig. 6.
Since the auto-negotiation function is disabled on the relay GbE line as
mentioned above, a line fault in the forward direction, if any, would not
cause
down-link of the same section in the backward direction.
As illustrated in Fig. 11, a fault occurs on a relay GbE line which
interconnects FE multiplexer 6-1 and GbE multiplexer 5, and as link-down is
accordingly detected by GbE MAC processing unit 20-1 in GbE multiplexer 5,
43
CA 02519617 2005-09-14
GbE multiplexer 5 assumes a relay line fault in the forward direction. Then,
if
the relay GbE line has not been recovered from link-down even after the lapse
of APS protection time TAPS, FE MAC processing circuit 16-1 in FE multiplexer
6-1 forcefully transitions a client line downstream of the failed relay GbE
line to
link-down to accomplish the I in k-pass-th rough for the downstream. This
processing is performed on every FE path which passes along the relay GbE
line on which the line fault has occurred.
Then, GBP processing circuit 17-1 sets a code ("0") indicative of the
absence of a line fault in the forward line fault notification field in the
GBP
transport header of a corresponding GBP capsule, a code (" 1") indicative of
the
presence of a line fault in the backward relay line fault notification field,
a code
indicative of no client line fault in the type field in the GBP core block,
and
allows the fixed length payload to pass therethrough as it is. The resulting
multiplexed MAC frame is sent in the backward direction. This processing is
performed on every FE path which causes the forced link-down to downstream
client lines.
The multiplexed MAC frame sent from FE multiplexer 6-1 is transferred
down to opposing FE multiplexer 2-1 which detects the backward relay line
fault
notification in the GBP transport header. FE multiplexer 2-1 immediately
transitions only a client line downstream of FE MAC processing circuit 16-1 to
link-down in a forcible manner.
Since the auto-negotiation function is enabled in a client line section as
mentioned above, FE paths in the forward direction also transition to link-
down.
Consequently, an alarm is additionally transferred to upstream client terminal
1-1-1 to accomplish the link-pass-through for the upstream client terminal.
This
processing is performed on every FE path which has undergone the
44
CA 02519617 2005-09-14
link-pass-through processing in the downstream direction resulting from the
fault on the relay GbE line.
Next, description will be given for an alarm transfer operation when a
fault occurs on a relay lOG SONET/SDH line.
Fig. 12 illustrates the alarm transfer operation when a fault occurs on a
relay 10G SONET/SDH line in the wide area Ethernet network illustrated in Fig.
6. Specifically, Fig. 12 illustrates the alarm transfer operation when a fault
occurs on a relay 10G SONET/SDH line which interconnects GbE multiplexer 3
and SONET/SDH cross-connect device 4.
As illustrated in Fig. 12, as a fault occurs on the relay lOG SONET/SDH
line, SONET/SDH cross-connect device 4 which detects the fault issues path
AIS alarm 24 to associated FE paths in the downstream direction. Specifically,
SONET/SDH cross-connect device 4 sets all SONET/SDH pointer values to "1"
and entirely overwrites a payload field with all "1. " This processing is
performed
only on an FE path on which a path AIS is detected when it is triggered by a
path fault which may occur in VC-4 units, and is performed on every FE path
which passes along the 10G SONET/SDH line when it is triggered by a link fault
such as an interruption of optical input.
Upon detection of the path AIS, 10G SONET generator/terminator circuit
23 in GbE multiplexer 5 further transfers the path AIS in the downstream
direction. Then, GBP relay processing circuit 21-1 sets a code (" 1")
indicative
of the presence of a line fault in the forward relay line fault notification
field in
the GBP transport header of a corresponding GBP capsule, a code ("0")
indicative of the absence of a line fault in the backward relay line fault
notification field, a code indicative of no client line fault in the type
field in the
GBP core block, and sets the fixed-length payload to a predefined pattern
(entirely comprised of " 1 " or the like) indicative of an idle frame.
CA 02519617 2005-09-14
The multiplexed MAC frame including this GBP capsule is transferred
down to GBP processing circuit 17-1 in FE multiplexer 6-1 which detects the
forward relay line fault notification in the GBP transport header. If the
forward
relay line fault has not been recovered even after the lapse of APS protection
time TAPS, FE multiplexer 6-1 forcefully transitions only a client line
downstream
of FE MAC processing circuit 16-1 to link-down to accomplish the
link-pass-through for the downstream. This processing is performed on every
FE path which passes along the relay 10G SONET/SDH line on which the fault
has occurred.
Then, GBP processing circuit 17-1 in FE multiplexer 6-1 sets a code ("0")
indicative of the absence of a line fault in the forward line fault
notification field
in the GBP transport header of a corresponding GBP capsule, a code ("1 ")
indicative of the presence of a line fault in the backward relay line fault
notification field, a code indicative of no client line fault in the type
field in the
GBP core block, and allows the fixed-length payload to pass therethrough as it
is. The resuiting multiplexed MAC frame is sent in the backward direction.
The multiplexed MAC frame is transferred down to opposing FE
multiplexer 2-1 which detects the backward relay line fault notification in
the
GBP transport header. FE multiplexer 2-1 immediately transitions only a client
line downstream of FE MAC processing circuit 16-1 to link-down in a forcible
manner. Since the auto-negotiation function is enabled in a client line
section,
FE paths in the forward direction also transitioned to link-down.
Consequently,
an alarm is additionally transferred to an upstream client terminal 1-1-1 to
accomplish the link-pass-through. This processing is performed on every FE
path which has undergone the link-pass-through processing in the downstream
direction resulting from the fault on the relay lOG SONET/SDH line.
46
CA 02519617 2005-09-14
In the configuration illustrated in Fig. 6, frames from a plurality of client
lines are multiplexed for transfer to a relay GbE line, they are again
multiplexed
for transfer to a relay 10G SONET/SDH line, frames passing through the 10G
SONET/SDH network are demultiplexed for transfer to relay GbE lines, and
they are again demultiplexed for transfer to client lines. Alternatively, for
example, frames demultiplexed by the GbE multiplexer may be again
multiplexed for transfer to a relay 10G SONET/SDH line, and the multiplexed
frames may be again demultiplexed for transfer to relay GbE lines, and they
may be again demultiplexed for transfer to client lines. Alternatively, frames
from a plurality of client iines may be transferred only through the relay 10G
SONET/SDH lines. As such, the first embodiment can be applied to other
network configurations as well.
Also, Fig. 6 illustrates the configuration in which client lines are provided
by an FE network, and relay lines are provided by a GbE network and a 10G
SONET/SDH network. Alternatively, the first embodiment can also be applied
to a network configuration in which client lines are provided by a GbE
network,
they are multiplexed for transfer to 10-gigabit Ethernet (10GbE) lines, they
are
again multiplexed for transfer to 40G SONET/SDH lines, frames which have
passed through a 40G SONET/SDH network are demultiplexed for transfer to
10-gigabit Ethernet (10GbE) lines, and they are again demultiplexed for
transfer
to client lines provided by a GbE network.
According to the alarm transfer method and wide area Ethernet network
of the invention, which is claimed in parent Canadian Patent Application
Serial
No. 2,458,694, the GBP transport header of the GBP capsule is provided with
the forward line fault notification field and backward relay line fault
notification
field, so that a notification on a fault occurring on a relay line can be
transferred
in the forward direction and backward direction, respectively. Further, since
a
47
CA 02519617 2005-09-14
{
fault on a client line can be notified using the type field in the GBP core
block,
the link-pass-through can be accomplished for downstream client terminals in
Ethernet path units at the egress node, while the link-pass-through can be
accomplished for upstream client terminals in Ethernet path units at the
ingress
node.
In addition, when a forward relay line fault notification is transferred
down to the egress node, while a backward relay line fault notification is
transferred from the egress node to the ingress node, information on a fault
occurring on a relay GbE line, a relay 10G SONET/SDH line and the like can be
transferred to a terminal of a communication partner, thereby making it
possible
to accomplish the link-pass-through of a relay line fault even in a wide area
Ethernet network configuration which utilizes a plurality of types of
transmission
networks.
Furthermore, the forward relay line fault notification is transferred down
to the egress node, from which a backward relay line fault notification is
issued,
rather than being returned from an intervening relay device as a backward
relay
line fault notification, so that the APS timer circuit need not be provided in
the
relay devices but is required only in the egress node. The APS timer circuit
counts the protection time which is required for switching one transmission
path
to another as mentioned above, and switches a transmission path of a relay
line when the forward relay line fault notification is continuously received
even
after the lapse of the time TAPs defined by the APS timer circuit. Generally,
the
APS timer circuit must be provided in each Ethernet path, so that the APS
timer
circuit in each of GbE multiplexers, mxn APS timer circuits are required, for
example, in the network configuration illustrated in Fig. 6. On the other
hand,
since the APS timer circuit is provided in each FE multiplexer in the first
embodiment, a required quantity of the APS timer circuits is only m which is
48
CA 02519617 2005-09-14
equal to the number of FE multiplexers installed in the network configuration
of
Fig. 6. Consequently, it is possible to limit the circuit scale per node.
Second Embodiment:
A wide area Ethernet network according to a second embodiment
provides redundancy for relay sections thereof. For this purpose, the second
embodiment proposes a mechanism which relies on fault notification and the
like transferred by the alarm transfer method shown in the first embodiment to
switch a relay line from a normally used active route to a spare route, and
vice
versa.
Fig. 13 illustrates the configuration of the wide area Ethernet network
according to the second embodiment, and Fig. 14 illustrates the configuration
of an FE multiplexer shown in Fig. 13. Fig. 15 shows exemplary classifications
for the operation of an alarm processing circuit contained in the FE
multiplexer
illustrated in Fig. 14.
In the wide area Ethernet network of the second embodiment, a plurality
of Ethernet networks, each of which accommodates a plurality of client
terminals, are relayed by a GbE network which comprises relay lines or GbE
lines that are made redundant. The Ethernet networks which accommodate
client terminals are implemented by FE networks which provide the
transmission capability of 10 Mbps or 100 Mbps. It should be noted that GBP
(Generic Blocking Procedure) is utilized for encapsulating higher-level
protocol
data such as a MAC frame and the like delivered from the client terminal, as
is
the case with the first embodiment. Since the frame format for GBP is similar
to that in the first embodiment, description thereon is omitted here.
As illustrated in Fig. 13, the wide area Ethernet network of the second
embodiment comprises FE multiplexers 102, 103, which are edge nodes,
49
CA 02519617 2005-09-14
interconnected through two redundant GbE lines #1, #2. Client terminals 101-1
to 101-m are connected to FE multiplexer 102 which is one of the edge nodes,
while opposing client terminals 104-1 to 104-m are connected to FE multiplexer
103 which is the other edge node. One of GbE line #1 and GbE line #2 is
assigned as an active route, and the other as a spare route. Client terminals
101-1 to 101-m are identical in configuration to client terminals 104-1 to 104-
m,
though they are installed at locations opposite to each other across relay
lines.
The client terminals may be Ethernet switches such as hubs.
In the following, each line for interconnecting a client terminal and an FE
multiplexer shown in Fig. 13 is called the "client line," and a line for
interconnecting FE multiplexers is called the "relay GbE line. " The number m
of
multiplexing provided by respective FE multiplexers 102, 103 is set, for
example, to a positive integer equal to or less than eight from the
relationship
between the capacity of the GbE line and an FE frame transfer band. In other
words, the number m of FE paths equal to eight or less are accommodated in a
multiplexed MAC frame (see Figs. 4A to 4C) which is transferred between FE
multiplexers 102, 103 through relay GbE lines.
For initiating a communication between client terminals opposing each
other across the relay GbE lines in the wide area Ethernet network illustrated
in
Fig. 13, as a source client terminal and a destination client terminal are
determined, multiplexing/demuttiplexing orders are determined in associated
FE multiplexers in accordance with a transmission path which interconnects
these client terminals. By thus determining line settings in each node in
accordance with the source client terminal and destination client terminal,
higher-level protocol data sent from an arbitrary client terminal is
transferred
only to a determined client terminal. A data flow which passes through a fixed
transmission path built by the line settings for transmitting Ethernet frames
is
CA 02519617 2005-09-14
called the "Ethernet path. " Here, such a transmission path is called the "FE
path" when the client line is an FE line, and is called the "GbE path" when
the
client line is a GbE path. In Fig. 13 assume that an FE path is set for a
transmission path from client terminal 101-1 to client terminal 104-1, and
likewise, FE paths are set between client terminals 101-2 and 104-2; between
client terminals 101-3 and 104-3; . . . ; between client terminals 101-m and
104-m, respectively. An FE path may be set between arbitrary client terminals.
Next, the FE multiplexer in the second embodiment will be described in
detail with reference to Fig. 14. As illustrated in Fig. 14, FE multiplexer
102
comprises FE PHY processing circuits 105-1 to 105-m each for
transmitting/receiving FE frames to/from an associated client line; FE MAC
processing circuits 106-1 to 106-m each -for performing MAC layer processing
for the FE frame; GBP generator/terminator circuits 107-1 to 107-m each for
encapsulating FE frames received from each FE line into GBP capsules and
decapsulating GBP capsules into FE frames for transmission onto each FE
line; multiplexer circuit 108 for multiplexing/demultiplexing GBP capsules of
each FE path in a predefined order; switching unit 109 for switching GbE line
#1 and GbE line #2, which are redundant relay GbE lines, from one to the other
as the active route and spare route; multiplexed MAC frame header
generator/terminator circuits 110-1, 110-2 each for generating and terminating
a multiplexed MAC frame header for each GbE line; GbE MAC processing
circuits 111-1, 111-2 each for adding a normal GbE MAC header and FCS to
data to which the multiplexed MAC frame header has been added; GbE PHY
processing circuits 112-1, 112-2 each for transmitting/receiving GbE frames
to/from the relay GbE lines; alarm processing circuit 114 for generating an SD
(Signal Degrade) signal and an SF (Signal Fail) signal for switching GbE lines
#1, #2; and APS (Automatic Protection Switching) processing circuit 113 for
51
CA 02519617 2005-09-14
determining line situations on the active route and spare route of the relay
GbE
lines to select the active route or spare route. FE multiplexer 103 is similar
to
FE multiplexer 102 in configuration.
FE PHY processing circuits 105-1 to 105-m each comprise a physical
device for transmitting FE frames, and a physical device for receiving FE
frames, respectively, for transmitting/receiving FE frames to/from an
associated
client line. FE PHY processing circuits 105-1 to 105-m also have a function of
detecting an interruption of FE input.
FE MAC processing circuits 106-1 to 106-m each perform the MAC layer
processing on the FE frames, and have a function of detecting FE link-down.
GBP generator/terminator circuits 107-1 to 107-m each encapsulate FE
frames received from each FE line into GBP capsules and decapsulate GBP
capsules into FE frames, and have a CRC error detecting function for the GBP
capsule, and an alarm transfer function.
Multiplexer circuit 108 multiplexes or demultiplexes GBP capsules for
each FE path in a predefined order.
Switching unit 109 switches GbE line #1 and GbE line #2, which are
redundant relay GbE lines, to the active route or spare route in response to a
switching signal from APS processing circuit 113.
Multiplexed MAC frame header generator/terminator circuits 110-1,
110-2 each generate and terminate a multiplexed MAC frame header for each
GbE line. Specifically, each multiplexed MAC frame header
generator/terminator circuit 110-1, 110-2 generates a sequence number, a K1
byte, and a K2 byte and performs HEC operation processing on the
transmission side, and checks the continuity of sequence numbers, extracts the
K1 byte and K2 byte, and checks the HEC operation on the reception side.
52
CA 02519617 2005-09-14
Each multiplexed MAC frame header generator/terminator circuit 110-1, 110-2
also has an APS byte defect detecting function.
GbE MAC processing circuits 111-1, 111-2 each add a MAC header of
normal GbE and FCS to data to which the multiplexed MAC frame header has
been added. GbE MAC processing circuit 111-1, 111-2 also have a function of
detecting link-down on a GbE line, and an FCS error detecting function on the
reception side.
GbE PHY processing circuits 112-1, 112-2 each comprise a physical
device for transmitting GbE frames, and a physical device for receiving GbE
frames, and transmit/receive GbE frames to/from relay GbE lines. GbE PHY
processing circuit 112-1, 112-2 additionally have a function of detecting an
interruption of GbE optical input.
Alarm processing circuit 114 collects a variety of detected error signals
from the respective circuits mentioned above to generate the SD signal and SF
signal for switching GbE lines #1, #2, and applies. APS processing circuit 113
with the SD signal and SF signal for each of GbE lines #1, #2.
APS processing circuit 113 uses the SD signal and SF signal for each of
GbE lines #1, #2 supplied from alarm processing circuit 114, and the
reception-related K1 byte and reception-related K2 byte supplied from
multiplexed MAC frame header generator/terminator circuits 110-1, 110-2 to
determine the line situations for the active route and spare route,
respectively,
to select between the active route or spare route. A selection signal
indicative
of the result of the selection is transferred to switching unit 109. APS
processing circuit 113 also generates a transmission-related K1 byte and a
transmission-related K2 byte which are delivered to multiplexed MAC frame
header generator/terminator circuits 110-1, 110-2. The selection between the
active route or spare route, and the generation of the K1 byte and K2 byte are
53
CA 02519617 2005-09-14
both performed in accordance with ITU-T Recommendation G. 841 (October
1998). In a configuration in which no redundant relay GbE line is provided,
the
relay GbE line switching signal is not changed and the value is taken by the
SD
signal and SF signal for each GbE line, and the reception-related K1 byte and
reception-related K2 byte.
Fig. 15 illustrates an alarm identifying method for each error detection
signal, which is executed by alarm processing circuit 114 in the second
embodiment. As illustrated in Fig. 15, alarms detected by alarm processing
circuit 114 are classified into an alarm detected for each FE path, and an
alarm
detected for each GbE line (each section). These detected alarms may be
delivered to OpS (network operation system). Also, upon detection of an alarm
classified as SD and an alarm classified as SF from the detected alarms, alarm
processing circuit 114 applies APS processing circuit 13 with a signal
indicative
of SD or SF in units of GbE lines (sections).
While Fig. 15 shows the alarm identification on the assumption that this
is applied in a network configuration which is provided with redundant relay
GbE lines, alarm processing circuit 114 can make the alarm identification
shown in Fig. 15 from the results of error detections in the respective
circuits
shown in Fig. 13, even in a network configuration which is not provided with
redundant relay GbE lines.
Next, details will be provided on the operation of selecting one from the
redundant relay GbE lines. The following description will be made for an
example in which GbE line #1 is assigned as the active route, and GbE line #2
as the spare route in the operation of the wide area Ethernet network
illustrated
in Fig. 13.
In these settings, FE multiplexer 102 on the transmission side generates
a multiplexed MAC frame from data received from client terminals 101-1 to
54
CA 02519617 2005-09-14
101-m, and applies the multiplexed MAC frame to GbE line #1 and GbE line #2,
respectively. FE multiplexer 103 on the reception side receives multiplexed
MAC frames from both active and spare relay lines. Then, FE multiplexer 103
extracts main signal data from GbE line #1 which is the active route, extracts
the K1 byte and K2 byte from GbE line #2 which is the spare route, and
determines in APS processing circuit 113 in accordance with the extracted
reception-related K1 byte and K2 byte whether or not the switching should be
made.
In this event, if a line fault occurs on GbE line #1 which is the active
route, a system switching operation is immediately performed by monitoring the
K1 byte and K2 byte communicated between FE multiplexers 102, 103, thereby
switching GbE line #2 to the active route and GbE line #1 to the spare route.
These operations can be similarly carried out when the active route is
initially
assigned to GbE line #2, and the spare route to GbE line #1.
In the wide area Ethernet network illustrated in Fig. 13, the client
terminals are implemented by FE terminals by way of example. Alternatively,
the wide area Ethernet network of the second embodiment can be applied to
any other higher-level protocol such as Fiber Channel as well.
According to the wide area Ethernet network of the second embodiment,
the transmission medium network layer is monitored for a fault using the
result
of the operation on the FCS field in the multiplexed MAC frame, and the path
network layer is monitored for a fault using the result of the operation on
the
CRC field in the GBP capsule for each Ethernet path. Consequently, the
section and path can be managed independently of one another as in the
SONET/SDH network, thereby making it possible to organize a transmission
network in a layered structure comprised of a transmission medium network
layer and a network layer.
CA 02519617 2005-09-14
Therefore, even in a wide area Ethernet network which applies GBP
encapsulation using an Ethernet network as an intermediate section, the
designing, maintenance, and operation of the network can be managed in a
layered structure, as in a wide area Ethernet network which uses a
SONET/SDH network as an intermediate section, to provide advanced network
services.
Third Embodiment:
A wide area Ethernet network according to a third embodiment provides
redundancy for a relay section thereof. For this purpose, the third embodiment
proposes a mechanism which relies on fault notification and the like
transferred
by the alarm transfer method shown in the first embodiment to switch a relay
line from a normally used active route to a spare route, and vice versa.
Fig. 16 illustrates the configuration of the wide area Ethernet network
according to the third embodiment. Fig. 17 illustrates the configuration of a
GbE multiplexer shown in Fig. 16 and Fig. 18 shows exemplary classifications
for the operation of an alarm processing circuit contained in the GbE
multiplexer illustrated in Fig. 17.
As illustrated in Fig. 16, the wide area network according to the third
embodiment comprises a plurality of Ethernet networks, each of which
accommodates a plurality of client terminals. The Ethernet networks are
accommodated in relay lines which make up a GbE network, and the GbE lines
are further multiplexed on and accommodated in a 10G SONET/SDH line which
has a transmission rate of 9. 953 Gbps. The Ethemet networks which
accommodate client terminals are implemented by FE networks which provide
the transmission capability of 10 Mbps or 100 Mbps. It should be noted that
GBP (Generic Blocking Procedure) is utilized for encapsulating higher-level
56
CA 02519617 2005-09-14
protocol data such as a MAC frame and the like delivered from the client
terminal, as is the case with the first embodiment. Since the frame format for
GBP is similar to that in the first embodiment, description thereon is omitted
here.
As illustrated in Fig. 16, the wide area Ethernet network of the third
embodiment comprises FE multiplexers 116-1 to 116-n, which are edge nodes,
are respectively connected through two redundant GbE lines #16-k-1, #1 6-k-2
(kis a positive integer which satisfies 1:4s n). Likewise, FE multiplexers 121-
1
to 121-n, which are edge nodes, are respectively connected through two
redundant GbE lines #21-k 1, #21-k-2 (k is a positive integer which satisfies
1 _<k<_n).
One of GbE line #16-k-1 and GbE line #16-k-2 is assigned as an active
route, and the other as a spare route. Likewise, one of GbE line #21-k-1 and
GbE line #21 -k-2 is assigned as an active route, and the other as a spare
route.
Client terminals 115-1-1 to 115-n-m are respectively connected to FE
multiplexers 116-1 to 11 6-n which are edge nodes, while opposing client
terminals 122-1-1 to 122-n-m are respectively connected to FE multiplexers
121-1 to 121 -n which are edge nodes. Client terminals 115-1-1 to 115-n-m are
identical in configuration to client terminals 122-1-1 to 122-n-m, although
they
are installed at locations opposite to each other across reiay lines. The
client
terminals may be Ethernet switches such as hubs.
In the wide area Ethernet network of the third embodiment, GbE
multiplexer 117 and Gbe multiplexer 120, each of which accommodates a
plurality of GbE lines, are connected to each other through SONET/SDH
cross-connect device 118 by way of redundant 10G SONET/SDH lines
(hereinafter also called the "relay SONET/SDH lines") #17-1, #17-2, #20-1,
#20-2.
57
CA 02519617 2005-09-14
SONET/SDH cross-connect device 118 has a function of
generating/terminating SOH of the SONET/SDH lines, and a function of
switching the redundant system based on the K1 byte and K2 byte. While
SONET/SDH cross-connect device 118 is connected to other lines as well to
provide appropriate cross-connect processing which supports respective lines,
it is assumed herein that SONET/SDH is set such that frames on 10G
SONET/SDH line #17-1 are delivered as they are to 10G SONET/SDH line
20-1, and frames on lOG SONET/SDH line #17-2 are delivered as they are to
lOG SONET/SDH line #20-2.
FE multiplexers 116-1 to 116-n, 121-1 to 121-n each have the same
functions and configuration as those in the second embodiment illustrated in
Figs. 13 and 14. The number m of multiplexing provided by respective FE
multiplexers 116-1 to 116-n, 121-1 to 121-n is set, for example, to a positive
integer equal to or less than eight from the relationship between the capacity
of
the GbE line and an FE frame transfer band. In other words, the number m of
FE paths equal to eight or less are multiplexed on a multiplexed MAC frame
(see Figs. 4A to 4C) which is transferred through the relay GbE line sections.
Likewise, the number m of multiplexing provided by respective GbE
multiplexers 117, 120 is set, for example, to a positive integer equal to or
less
than eight from the reiationship between the capacity of the 10G SONET/SDH
line and a band in which the data field in the multiplexed MAC frame is
transferred. When represented by the number of multiplexing on an FE path,
the number m is converted to a positive integer equal to or less than 64. In
other words, in the 10G SONET/SDH frame in the relay SONET/SDH line
section (see Figs. 5A to 5C), one FE path is transferred by VC-4, and the
number m of FE paths equal to 64 or less are multiplexed on a lOG
SONET/SDH frame.
58
CA 02519617 2005-09-14
For making a communication between client terminals opposing each
other across the relay GbE lines and relay SONET/SDH lines in the wide area
Ethernet network illustrated in Fig. 16, as a source client terminal and a
destination client terminal are determined, multiplexing/demultiplexing orders
are determined in associated FE multiplexers and GbE multiplexers in
accordance with transmission paths which connect these client terminals,
followed by determination of output ports on the SONET/SDH cross-connect
device 118. By thus determining line settings in each node in accordance with
the source client terminal and destination client terminal, higher-level
protocol
data sent from an arbitrary client terminal is transferred only to a
determined
client terminal. Data flow which passes through a fixed transmission path
built
by the line settings for transmitting Ethemet frames is caited the "Ethernet
path.
" Here, such a transmission path is called the "FE path" when the client line
is
an FE line, and is called the "GbE path" when the client line is a GbE path.
In
Fig. 16 assume that an FE path is set for a transmission path from client
terminal 115-1-1 to client terminal 122-1-1, and likewise, FE paths are set
between client terminals 115-1-2 and 122-1-2; between client terminals 115-1-3
and 122-1-3; . . . ; between client terminals 115-n-m and 122-n-m,
respectively. An FE path may be set between arbitrary client terminals.
Next, the GbE multiplexer in the third embodiment will be described with
reference to Fig. 17. As illustrated in Fig. 17, GbE multiplexer 117 comprises
GbE PHY processing circuits 123-1-1 to 123-n-1, 123-1-2 to 123-n-2 each for
transmitting/receiving multiplexed MAC frames to/from an associated relay GbE
line; multiplexed MAC frame header generator/terminator circuits 125-1-1 to
125-n-1, 125-1-2 to 125-n-2 each for generating and terminating a multiplexed
MAC frame header for each GbE line; GBP relay processing circuits 126-1-1 to
126-n-1, 126-1-2 to 126-n-2 each for detecting CRC errors in GBP capsules
59
CA 02519617 2005-09-14
and transferring an alarm; multiplexer circuits 127-1, 127-2 each for
multiplexing/demultiplexing GBP capsules of each FE path extracted from
multiplexed MAC frames in a predefined order; switching unit 128 for switching
redundant relay GbE lines and relay SONET/SDH lines; lOG PHY processing
circuits 129-1, 129-2 each for transmitting/receiving 10G SONET/SDH frames
to/from an associated SONET/SDH line; SONET/SDH transmission/reception
processing circuits 130-1, 130-2 each for generating values for respective
bytes
in the SOH field and POH field for a signal in the SONET/SDH line format on
the transmission side, and for detecting each section alarm and each path
alarm defined by SONET/SDH from the SOH field and POH field on the
reception side; alarm processing circuit 131 for generating an SD (Signal
Degrade) signal and an SF (Signal Fail) signal for switching a GbE line or a
lOG SONET/SDH line; and APS (Automatic Protection Switching) processing
circuit 132 for determining line situations on the active route and spare
route of
the relay GbE lines to select the active route or spare route. GbE multiplexer
120 (shown in Fig. 6) is similar to GbE multiplexer 117 in configuration
described above.
GbE PHY processing circuits 123-1-1 to 123-n-1, 123-1-2 to 123-n-2
each comprise a physical device for transmitting GbE frames, and a physical
device for receiving GbE frames, and transmit/receive multiplexed MAC frames
to/from an associated reiay GbE line. GbE PHY processing circuits 123-1-1 to
123-n-1, 123-1-2 to 123-n-2 also have a function of detecting an interruption
of
GbE input.
GbE MAC processing circuits 124-1-1 to 124-n-1, 124-1-2 to 124-n-2
each perform MAC layer processing on the FE frames, and has a function of
detecting FE link-down.
CA 02519617 2005-09-14
Multiplexed MAC frame header generator/terminator circuits 125-1 -1 to
125-n-1, 125-1-2 to 125-n-2 each generate and terminate a multiplexed MAC
frame header for each GbE line. Specifically, each multiplexed MAC frame
header generator/terminator circuit 125-1-1 to 125-n-1, 125-1-2 to 125-n-2
generates a sequence number, a K1 byte, and a K2 byte, and performs HEC
operation processing on the transmission side, and checks the continuity of
sequence numbers, extracts the K1 byte and K2 byte, and checks the HEC
operation on the reception side. Multiplexed MAC frame header
generator/terminator circuits 125-1-1 to 125-n-1, 125-1-2 to 125-n-2 also have
an APS byte defect detecting function.
GBP relay processing circuits 126-1-1 to 126-n-1,126-1-2 to 126-n-2
each monitor GBP capsules in multiplexed MAC frames transmitted through
each GbE line to detect CRC errors in the GBP capsules, and transfer an
alarm. Also, GBP relay processing circuits 126-1-1 to 126-n-1,126-1-2 to
126-n-2 have a function of inserting an idle frame defined by the GBP capsule
on the transmission side, and a function of extracting the idle frame defined
by
the GBP capsule on the transmission side in order to adjust a difference in
rate
with the FE frame when data is stored in the payload of a VC-4 frame by
multiplexer circuits 127-1,127-2.
Multiplexer circuits 127-1,127-2 each multiplex/demultiplex GBP
capsules of each FE path extracted from multiplexed MAC frames in a
predefined order. A signal delivered from multiplexer circuits 127-1,127-2 to
switching unit 128 is defined by the format shown in Fig. 5C, and data on one
FE path is stored in one VC-4 frame. However, the format shown in Fig. 5C
only reserves fields for the POH and SOH signals, the values of which are
generated by SONET/SDH transmission/reception processing circuits 130-1,
130-2.
61
CA 02519617 2005-09-14
Switching unit 128 switches the redundant GbE lines between the active
route and spare route in response to a switching signal from APS processing
circuit 132. The relay GbE lines are redundantly organized by GbE line #16-k-1
and GbE line #16-k-2 which are set to transfer GBP blocks in an FE path from
the active GbE line to a 10G SONET/SDH line. Likewise, the relay
SONET/SDH lines are redundantly organized, and are set to transfer GBP
blocks in an FE path from the active 10G SONET/SDH line to a GbE line.
lOG PHY processing circuits 129-1, 129-2 each comprise a physical
device for transmitting 10G SONET/SDH frames, a physical device for receiving
10G SONET/SDH frames, and transmit/receive 10G SONET/SDH frames
to/from the relay SONET/SDH lines. lOG PHY processing circuits 129-1, 129-2
also have a function of detecting an interruption of 10G SONET/SDH optical
input.
SONET/SDH transmission/reception processing circuits 130-1, 130-2
each generate a value for each of the bytes in SOH and POH for multiplexed
signals in the 10G SONET/SDH line format transferred from switching unit 128,
and insert the generated values in appropriate fields on the transmission
side.
On the reception side, SONET/SDH transmission/reception processing circuits
130-1, 130-2 each detect each section alarm and each path alarm defined by
SONET/SDH respectively from the SOH field and POH field in each received
frame. The detected section alarm is relayed to alarm processing circuit 131
on
a line-by-line basis, while the path alarm is notified for each path of each
line.
Alarm processing circuit 131 collects a variety of error signals detected
by the respective circuits mentioned above to generate the SD signal and SF
signal for switching the GbE line or 10G SONET/SDH line, and applies APS
processing circuit 132 with the SD signal and SF signal for each GbE line or
for
each lOG SONET/SDH line.
62
CA 02519617 2005-09-14
APS processing circuit 132 uses the SD signal and SF signal for each
GbE line supplied from alarm processing circuit 131, and the reception-related
K1 byte and reception-related K2 byte supplied from multiplexed MAC frame
header generator/terminator circuits 125-1-1 to 125-n-1,125-1-2 to 125-n-2 to
determine the line situations for the active route and spare route of the
relay
GbE lines, respectively, to select the active route or spare route. A
selection
signal is transferred to switching unit 128. Likewise, APS processing circuit
132
determines the line situations for the active route and spare route of the
relay
SONET/SDH lines, respectively, from the SD signal and SF signal for each 10G
SONET/SDH line supplied from alarm processing circuit 131, and the
reception-related K1 byte and reception-related K2 byte supplied from
SONET/SDH transmission/reception processing circuits 130-1, 130-2 to select
the active route or spare route. A selection signal is transferred to
switching
unit 128.
APS processing circuit 132 also generates a transmission-related K1
byte and a transmission-related K2 byte which are delivered to multiplexed
MAC frame header generator/terminator circuits 125-1-1 to 125-n-1,125-1-2 to
125-n-2 . The selection of the active route or spare route, and the generation
of the K1 byte and K2 byte are both performed in accordance with ITU-T
Recommendation G. 841 (October 1998).
In a configuration in which no redundant relay GbE line is provided, the
relay GbE line switching signal should not be changed and the value is taken
by
the SD signal and SF signal for each GbE line, and the reception-related K1
byte and reception-related K2 byte. Likewise, in a configuration in which no
redundant SONET/SDH line is provided, the relay SONET/SDH line switching
signal should not be changed and the value is taken by the SD signal and SF
63
CA 02519617 2005-09-14
signal for each SONET/SDH line, and the reception-related K1 byte and
reception-related K2 byte.
Fig. 18 shows an alarm identifying method for each error detection
signal, executed by the alarm processing circuit in the third embodiment. As
shown in Fig. 18, alarms detected by alarm processing circuit 131 are
classified
into an alarm detected for each FE path, an alarm detected for each GbE line
(each section), an alarm detected for each VC-4 path, and an alarm detected
for each 10G SONET/SDH line (each section). These detection alarms may be
delivered to OpS (network operation system). Also, upon detection of an alarm
classified as SD and an alarm classified as SF out of detected alarms, alarm
processing circuit 131 applies APS processing circuit 132 with a signal
indicative of SD or SF in units of GbE lines (sections) or in units of 10G
SONET/SDH lines (sections).
While Fig. 18 shows the alarm identification on the assumption that this
is applied in a network configuration which is provided with redundant relay
GbE lines and redundant relay SONET/SDH lines, alarm processing circuit 131
can make the alarm identification shown in Fig. 18 from the results of error
detections in the respective circuits shown in Fig. 16 even in a network
configuration which is not provided with redundant relay GbE lines or
redundant
relay SONET/SDH lines.
Next, description will be made on the operation of selecting one from the
redundant relay GbE lines and one from the redundant relay SONET/SDH
lines.
The following description will be made for an example in which GbE line
#16-1-1 is assigned as the active route, and GbE line #16-1-2 as the spare
route in the operation of the wide area Ethernet network illustrated in Fig.
16.
64
CA 02519617 2005-09-14
In these settings, FE multiplexer 116-1 on the transmission side
generates a multiplexed MAC frame from data received from client terminals
115-1-1 to 115-1 -m, and applies the multiplexed MAC frame to GbE line
#16-1-1 and GbE line #16-1-2, respectively.
FE multiplexer 117 receives multiplexed MAC frames from both active
and spare relay lines. Then, FE multiplexer 117 extracts main signal data from
GbE line #16-1-1 which is the active route, extracts the K1 byte and K2 byte
from GbE line #16-1-2 which is the spare route, and determines in APS
processing circuit 132 in accordance with the extracted reception-related K1
byte and K2 byte whether or not the switching should be made.
In this event, if a line fault occurs on GbE line #16-1-1 which is the active
route, a system switching operation is immediately performed by monitoring the
K1 byte and K2 byte communicated between FE multiplexers 116-1 and GbE
multiplexer 117, thereby switching GbE line #16-1-2 to the active route and
GbE line #16-1-1 to the spare route. These operations can be similarly carried
out when the active route is initially assigned to GbE line #16-1-2, and the
spare route to GbE line #16-1-1. Also, these operations can be similarly
carried
out in another redundant system of the relay GbE lines between FE
multiplexers 116-i (i is a positive integer satisfying 2<_i<_n) and GbE
multiplexer
117. Further, these operations can be similarly carried out in another
redundant system of the relay GbE lines between FE multiplexers 121-j(j is a
positive integer satisfying 2< j<n) and GbE multiplexer 120.
Next, description will be made for an operation in which 10G
SONET/SDH line #17-1 is assigned as the active route, and 10G SONET/SDH
line #17-2 as the spare route in the wide area Ethernet network.
CA 02519617 2005-09-14
In these settings, GbE multiplexer 117 delivers frames to active and
spare 10G SONET/SDH line #17-1 and 10G SONET/SDH line #17-2,
respectively.
SONET/SDH cross-connect device 118 receives frames from both the
active route and spare route. Then, SONET/SDH cross-connect device 118
extracts main signal data from 10G SONET/SDH line #17-1 which is the active
route, extracts the K1 byte and K2 byte from 10G SONET/SDH line #17-2
which is the spare route, and determines in accordance with the extracted
reception-related K1 byte and K2 byte whether or not the switching should be
made.
In this event, if a line fault occurs on 10G SONET/SDH line #17-1 which
is the active route, a system switching operation is immediately performed by
monitoring the K1 byte and K2 byte communicated between GbE multiplexer
117 and SONET/SDH cross-connect device 118, thereby switching 10G
SONET/SDH line #17-2 to the active route and lOG SONET/SDH line #17-1 to
the spare route. These operations can be similarly carried out when the active
route is initially assigned to 10G SONET/SDH line #17-2, and the spare route
to
10G SONET/SDH line #17-1. Also, the operations can be similarly carried out
for lines #20-1 and #20-2 between GbE multiplexer 120 and SONET/SDH
cross-connect device 118.
In the wide area Ethernet network illustrated in Fig. 16, the client
terminals are implemented by FE terminals by way of example. Altemativety,
the wide area Ethernet network of the third embodiment can be applied to any
other higher-level protocol such as Fiber Channel as well.
According to the wide area Ethernet network of the third embodiment,
even when a transmission network in a relay section is implemented by a
combination of an Ethernet (GbE) network and a SONET/SDH network such
66
CA 02519617 2005-09-14
that the SONET/SDH network is utilized as a transmission network in a certain
section, while the GbE network is utilized as a transmission network in
another
section, transmission medium network layers corresponding to respective
sections can be managed irrespective of the type of the transmission networks.
Since the GBP capsule can be used to manage a path network layer from an
ingress node to an egress node on an end-to-end basis, the networks can be
designed, operated, and maintained without knowledge of the difference
between the transmission networks, thereby providing advanced network
services. Furthermore, in such a configuration, it is possible to reduce the
circuit scale and mounting area of a relay node which corresponds to a
connection of a SONET/SDH network with an Ethernet network. This is a result
of the respective Ethernet networks being provided with the same switching
means as that use in the SONET/SDH network in order to share the APS
processing circuit provided for the SONET/SDH network. When a single APS
processing circuit can be shared for the SONET/SDH and Ethernet networks,
the relay node can be reduced in circuit scale and mounting area.
In addition, the automatic protection switching (APS) can be performed
as fast as in the SONET/SDH network by using K1 byte and K2 byte and
applying similar APS processing to that used in the SONET/SDH network in a
section in which the Ethernet is used as a transmission network.
While preferred embodiments of the present invention have been
described using specific terms, such description is for illustrative purposes
only,
and it is to be understood that changes and variations may be made without
departing from the spirit or scope of the following claims.
67
